- COPYDOT-CG-D-04-90 An ExperimentalDOT-TSC-CG-90-2 Evaluation of a Field

N Sobriety Test Batteryoo in the Marine Environment

NE. Donald SussmanAnn NeedalmanPeter H. Mengert

Transportation Systems CenterCambridge, MA 02142

June 1990Final Report

This document is available to the publicthrough the National Technical InformationService, Springfield, Virginia 22161

DTICELECTEOCT031

U.S. Departmentof TransporttionUnited StaesAfCoast Guard

Office of Navigation Safety and Waterway ServicesAuxiliary, Boating and Consumer Affairs DivisionWashington, DC 20593

90

NOT: CE

This document is disseminated under the sponsorshipof the Department of Transportation in the interestof information exchange. The United States Govern-ment assumes no liability for its contents or usethereof.

NOTICE

The United States Government does not endorse pro-ducts or manufacturers. Trade or manufacturers'names appear herein solely because they are con-sidered essential to the object of this report.

Technical Report Documentation Page

1. Re ort No. 2. Goernment Accession No. 3. Recipient's Catalog No.

D)OT-CG-D-04-90

4. Title and Subtitle 5. epor Date

An Experimental Evaluation of a Field Sobriety Test June 1990

Batter in the Marine Environment 6. Pertot.c ngOrganization CodeDTS-45

8. Performing Organization Report No.7. Author' s)

E. D~onald Sussman, Peter H. Mengert DOT-TSC-C(-90-2

9. Performing Organization Name and Address 10. Work Unit No. (TRAIS)

U.S. Department of Tro nsportaticn CGO74/BO024

Research and Special Programns Administration I I. Contract orGrantNo.

Transportation "vstems CenterCabridge, MiA 02142 13. Type of Report and Period Covered

12. Sponsoring Agency Name and Address Final ReportU.S. Department of Transportation March 1987-February 1990United States Coast GuardOffice of Navigation Safety and Waterway Services 14. Sponsoring Agency Code

Washington, DC 20593 G-NAB

15. Supplementary Notes

16. Abstract

This Report describes an investigation of the accuracy of a FST (Field Sobriety Test)battery used in the marine environment. FSTs rely on the observation and measurement ofthe effect of alcohol intoxication on coordination, visual tracking and balance. Thepurpose of this study was to determine if there was any decrease in the accuracy ofthe tests when used under recreational boating conditions.

In the study, 97 volunteers were dosed with alcohol (four drinks over three and onehalf hours) in a recreational boating setting. The subjects' BACs (Blood AlcoholConcentration) were estimated through FST orocedures by marine law enforcement agentsexperienced in the use of such procedures.

The officers correctly identified the subjects who would be legally intoxicated (BACsequal to or greater than 0.10%) in 82% of all cases.

The overall correlation of the officers' FST based estimates with BACs obtained usingbreath tests was approximately .70. This level is consistent with similarly obtainedcorrelations from highway studies.

Calculation of indices of the officers' arrest-release performances revealed that FSTtests used on the water can result in the arrest of significantly more intoxicatedboaters while maintaining a very low level of false arrests.

It was concluded that the accuracy of FST batteries are not degraded in the marineenvironment.

17. Key Words 18. Distribution Statement

Field Sobriety Tests, DOCUMENT IS AVAILABLE TO THE PUBLICTHROUGH THE NATIONAL TECHNICAL

Alcohol, Recreational Boats INFORMATION SERVICE. SPRINGFIELD.VIRGINIA 22161

19. Security Classif. (of this report) 20. Security ClasiI. (of this page) 21. No. of Pages 22. Price

UNCLASSIFIED UNCLASSIFIED 96

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

PrefaceThis study was carried out for the U.S. Coast Guard's Office of Boating, Public, andConsumer Affairs (after reorganization, the Office of Navigation Safety and Water-ways Services) and the U.S. Coast Guard's Office of Engineering, Logistics andDevelopment by the Transportation Systems Center's Operator Performance andSafety Analysis Division, in cooperation with the International Association of Chiefsof Police.

The study received the enthusiastic, creative, and active support of individuals from anumber of diverse organizations. We would like to acknowledge the efforts of thefollowing organizations and individuals:

The U.S. Coast Guard Headquarters staff members, who provided not only overallprogram direction and coordination but the encouragement and institutional supportwhich was vital to the execution of the study: A. J. Marmo, Capt. M.B. Stegner, Lt.John Smith, Lt. William Cairns, and CWO3 John Williams; the Yorktown ReserveTraining Center (RTC) staff who provided vital logistic support, In, particular Lt.Commander T Doherty, Lt. Mark Kern, and Chief "Corky" Elliot.

Capt. C. Richard Mockler, TSC Coast Guard Coordinator for his guidance and assis-tance.

The staff from the International Association of Chiefs of Police, Charles Peltier andRon Sostkowski for securing the cooperation of the police officers who served as ratersand for their aid in study planning and design. The creative and untiring efforts of theresearch team from Dunlap and Associates who recruited, dosed, and monitored thesubjects.

Robert DiSario and Paul Hoxie of TSC for their thorough and insightful review of thestatistical design and Arthur Flores and Arnold Spicer for their supervision of thebreath testing.

The members of the law enforcement community who were involved in the study; Capt.Frank Wood, Maryland Department of Natural Resources and Randy Smith and PamDillon, Ohio Department of Natural Resources for their support in providing the ratersand the six marine law enforcement officers who served as raters.

We also thank the ninety-seven men who served as subjects and their respectiveorganizations: RTC Yorktown, U.S. Coast Guard Fifth District, Fort Eustis, and theNaval Weapons Station. We also thank the volunteers from the U.S.Coast GuardAuxiliary who served as boat operators.

Finally special thanks is given to program monitor Jerry Boden for his continuedmeticulous and dedicated support of all phases of the effort from planning through finaltechnical review.

METRIC/ ENGLISH CONVERSION FACTORS

ENGLISH TO METRIC METRIC TO ENGLISH

LENGTH (APPROXI MATE) LENGTH (APPROXIMATE)

1 inch (in)= 2.5 centimeters (cm) 1 millimeter (mm) = 004 inch (in)

1 foot (ft) = 30 centimeters (cm) I centimeter (cm) = 0.4 inch (in)I yard (yd) = 0.9 meter (m) 1 meter (m) z 3.3 feet (ft)1 mile cmi) = 1.6 kilometers (km) 1 meter (m) = 1.1 yards (yd)

1 kilometer (kin) = 0.6 mile (mi)

AREA (APPROXIMATE) AREA (APPRCXWMATE)

I square inch (sq in, in-) = 6.5 square centimeters (cm 2) I square centimeter (cm-) = 0.16 square inch (sq in, in-)I square foot (sq ft, ft 2 ) = 0.09 square meter (m2 ) 1 square meter (m) = 1.2 square yards (sq yd, yd 2 )

1 square yard (sq yd, yd 2) = 0.8 square meter (M2) 1 square kilometer (kin2 ) = 0.4 square mile (sq mi, mi2)

1 square mile (sq mi, mi 2) = 2.6 square kilometers (kin2) 1 hectare (he) = 10,000 square meters (M2) = 2.5 acres

1 acre = 0.4 hectares (he) = 4,000 square meters (m2)

MASS - WEIGHT (APPROXWATE( MASS - WEIGHT wp~oxim:i1 ounce (oz) = 28 grams (gr) 1 gram (gr) = 0.036 ounce (oz)I pound (lb) = .45 kilogram (kg) 1 kilogram (kg) = 2.2 pounds (Ib)

1 short ton = 2,000 pounds (Ib) = 0.9 tonne (t) 1 tonne (t) = 1,000 kilograms (kg) = 1.1 short tons

VOLUME (APPROXIMA E) VOLUME (AoPoxiMATE)1 teaspoon (tsp) = 5 milliliters (ml) 1 milliliter (ml) = 0.03 fluid ounce (Il oz)

1 tablespoon (tbsp) = 15 milliliters (ml) I liter (1) = 2.1 pints (pt)1 fluidounce (fl oz) = 30 milliliters (ml) 1 liter (I) = 1.06 quarts (qt)

1 cup (c) = 0.24 liter (I) 1 liter (I) = 0.26 gallon (gal)1 pint (pt) = 0.47 liter (I) 1 cubic meter (m 3) = 36 cubic feet (cu ft ft.)

1 quart (qt) = 0.96 liter (1) 1 cubic meter (M3) = 1.3 cubic yards (cu yd, yd3)

1 gallon (gal) = 3.8 liters (I)1 cubic foot (cu ft, ft 3) = 0.03 cubic meter (M 3

)

1 cubic yard (cu yd, yd 3) = 0.76 cubic meter (m 3)

TEMPERATURE (EXACT) TEMPERATURE (EXAc)

[x -32)(5/9)J = y'C [(9/ 5 )y + 32] C = x'F

QUICK INCH-CENTIMETER LENGTH CONVERSION

INCHES 0 1 2 3 4 5 6 7 8 9 10SIIII!I I I

CENTIMETERS 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25I

25.40

QUICK FAHRENHEIT-CELCIUS TEMPERATURE CONVERSION'F -40' -22" -4" 14" 32' 50 68' 86' 104" 122' 140' 158' 176' 194' 212'

C C C I I I I C

1C -40. -30° -20° -10' 0. 10' 20' 30° 40' 50' 60' 70° 80 90' 100.

For more ewact and'or Other conversion factors. :ee NES Miscellaneous Publication 286. Units of Weights andMeasures. Price S2.50. SD Catalog No. C13 10286.

iv

Executive SummaryThis Report describes an investigation of the accuracy of a field sobriety test (FST)battery when used in the marine environment. FSTs are non-chemical tests of intoxi-cation which are used in highway law enforcement. These tests rely on the observationand measurement of the effect of alcohol intoxication on behaviors such as coordina-tion, visual tracking and balance. It has been suggested that such behaviors might bedegraded by the stressors encountered in the marine environment thereby invalidatingthem for such use. The purpose of this study was to determine if there was any decreasein the accuracy of the tests when used on individuals in recreational boating conditions.

In the study, 97 volunteers (who were similar in age to the population of individualsarrested for operating under intoxication (OUI) encountered in MD., OH., and twocounties in CA. during a two year period) were dosed with alcohol in a setting closelyapproximating that encountered in recreational boating. The subjects were given fourdrinks over a three and one half hour exposure period. The dosages were calculated tocause the subject to reach Blood Alcohol Concentrations (BACs) of 0.12%, 0.08%, or0%. During this exposure period the subjects spent approximately one and one halfhours on the water at various speeds in an open, high performance outboard type boat.

The subjects' BACs were estimated through FST procedures by rmarine law enforce-ment agents experienced in the use of such procedures. The FSTs were conducted bothon the water and on land. The officers correctly classified the subjects' BACs to beeither below 0.10% or equal to or greater than 0.10% in 82% of all cases. The CoastGuard's limit for OUI for recreational boating is 0.10%.

The officers' estimates were correlated with measurements of BAC obtained usingbreath testing units. The correlations obtained were similar to and consistent withcorrelations between FST estimates of BAC and breath test measurements found instudies conducted to simulate the highway environment. The overall correlation wasapproximately .70.

Indices of the officer's performance in correctly determining whether a subject did ordid not exceed an intoxication criteria were calculated. These indices revealed that,even when used in a "conservative" nanner, FST tests used on the water will result inthe arrest of a significantly greater number of intoxicated boaters than would bearrested using only observation and interrogation methods and would probably resultin a very low level of false arrests. ("Conservative" manner refers to a situation in whichthe officer would arrest only suspects who were believed to have BAC levels of at least0.02% above the legal BAC level for intoxication).

It was concluded that the accuracy of FSTs is not degraded by the marine environment.

V

TABLE OF CONTENTS

1 INTRO DUCTION .................................................................................................... 11.1 O bje ctiv e ............................................................................................................. 11.2 Study Background .......................................................................................... 21.3 Background - Field Sobriety Testing ......................................................... 2

.2 M ETHOD ............................................................................................................. 52.1 Design ............................................................................................................. 62.2 Field Sobriety Tests Used ........................................................................... 72.3 Subjects and Raters ....................................................................................... 9

2.3.1. Subject Description .................................................................... 92.3.2. Subject Screening ....................................................................... 102.3.3. Raters/M arine Law Enforcement Officers ........................... 11

2.4 Facilities and Equipment ........................................................................... 122.4.1. Study Site ..................................................................................... 122.4.2. Boats ............................................................................................ 122.4.3. Dosing Schedule and Apparatus ............................................ 122.4.4. Breath Testing Equipment ..................................................... 142.4.5. Routine and Emergency M edical Personnel ......................... 15

2.5 Preliminary Procedures ............................................................................. 152.5.1. Testing Schedule ...................................................................... 152.5.2. Environmental Conditions ..................................................... 152.5.3. Subject Briefing ......................................................................... 162.5.4. Officer/Rater Briefing ............................................................. 16

2.6 Study Procedures ....................................................................................... 172.6.1. Subject Arrival, Briefing, and Medical Examina-

tion ............................................................................................ 172.6.2. Dosing on Land ......................................................................... 182.6.3. Dosing on the Boat .................................................................... 192.6.4. Rendezvous with the Police Boat ............................................ 192.6.5. Testing BAC on Boat ............................................................... 202.6.6. Field Sobriety Testing on the Boat ........................................ 202.6.7. Field Sobriety Testing on Land ............................................... 242.6.8. M onitor and Release of Subject ............................................ 242.6.9. Officer Debrieting and Feedback Sessions .......................... 25

3 ANALYSIS ......................................................................................................... 26

vi

4 R E S U L T S ................................................................................................................. 294.1 Correlations Between Estimates of Bac and Measured BAC

...................................................................................................................... 2 94.1.1. Overall Correlation Between Bac Estimates Ac-

tual B A C .................................................................................. 294.1.2. Correlation Between Individual Rater's Estimates

and A ctual BA C s .................................................................... 294.2 Tests of Absolute Difference Between Estimated BACs and

actual B A C s .......................................................................................... 294.3 A rrest/R elease D ecision ....................................................................... 36

4.3.1. Criteria for Arrest at > =0.10% (equal to orgreater than 0.10%) Criteria for Conviction at> -- 0.10% ............................................................................... . . 37

4.3.2. Criteria for Arrest at > = 0.08% Criteria for Con-viction at> = 0.08% ............................................................... 44

4.3.3. Criteria for Arrest at > = 0.12% Criteria for Con-viction at > = 0.10% ............................................................. 44

4.3.4. Criteria for Arrest at > = 0.10% Criteria for Con-viction at > = 0.08% ............................................................. 47

4.3.5. Criteria for Arrest at > = 0.08% Criteria for Con-viction at > = 0.10% ............................................................. 47

4.4 Considering the Two-Phase Test Process .......................................... 50

5 C O N C LU SIO N S ................................................................................................. 53

6 RECOMMENDATIONS ................................................................................. 54

INDEX TO APPENDIX ITEMS ..................................................................... A- 1

A PPE N D IX ............................................................................................... A 2- A 30

R E F E R E N C E S ....................................................................................................... R -1

Accession For

NTIS GRA&I

DTIC TAB 0

cZ~ £~J ~ unannounced 0Q Justification

• _ " By_

Distribution/,

Availability Cod6e

Avail and/orDIet SpeG lal

vii I

LIST OF TABLES AND FIGURES

Figure Page

1. T he W aterfront A rea ........................................................................... 132. Position of Boats D uring Testing ........................................................ 193. Breath Testing Prior to Field Sobriety Testing ................................. 204. Blood Alcohol Concentration for Each Subject ............................... 21

Administering the HGN Test on the Boat ....................................... 236. Administering the Walk and Turn Test on Land ............................. 247. Correlation Between All Estimates and Measured BAC .............. 308. Correlation Between Individual Officer's Estimates and Mea-

sured B A C .......................................................................................... . . 329. Arrest Decision Point > = .10%, Convict Point > = . 10% .......... 4010. Arrest Decision Point > = .08%, Convict Point > = .08% ...... 4511. Arrest Decision Point > = .12%, Convict Point > = .10% ...... 4612. Arrest Decision Point > = .10%o, Convict Point > = .08% ...... 4813. Arrest Decision Point > = .08%, Convict Point > = .10% ...... 49

Table Page

I.BACs M easured Prior to FST ....................................................................... 222.Correlation Between Individual Rater's Estimates and BAC ................. 313.Absolute Difference Between Individual Rater's Estimates

and B A C ............................................................................................. . . 334a. Overall Arrest/Release Decision Estimates Arrest BAC > =

0.10%, Convict BAC > = 0.10% ...................................................... 384b. Overall Arrest/Release Decision Estimates Arrest BAC

> = 0.08%, Convict BAC > = 0.08% ............................................ 384c. Overall Arrest/Release Decision Estimates Arrest BAC > =

0.12%, Convict BAC > = 0.10% ...................................................... 384d. Overall Arrest/Release Decision Estimates Arrest BAC

> = 0.10%, Convict BAC > = 0.08% ............................................ 394e. Overall Arrest/Release Decision Estimates Arrest BAC > =

0.08%, Convict BAC > = 0.10% ...................................................... 395. Decision Index for All Raters ............................................................ 416. Comparison of Marine and Non-marine FST Indices .................... 427. A rrest E rror R ates ................................................................................ 438. The Two Stage Decision Process ........................................................ 52

viii

1 INTRODUCTION

1.1 Objective

The objective of this study was to determine the usefulness of a field sobriety test (FST)battery in assisting the marine law enforcement officer in determining if boat operatorsare intoxicated. This is part of the Coast Guard's effort to support local marine lawenforcement. This study was performed in cooperation with the International Associa-tion of Chiefs of Police (IACP), which will use the information generated in assemblinga boating safety manual focusing on the issue of alcohol detection.

Field sobriety tests have been verified as useful techniques in the detection of theintoxicated automobile operator (Tharp, et al., 1981). Since certain stressors arepresent in the boating environment which are not present on the highway, it is necessaryto evaluate the effectiveness of these tests in the marine environment. It has beenhypothesized that stressors inherent in marine operations (i.e., the effects of heat, spray,boat motion, vibration, glare) may cause boaters (whether intoxicated or sober) toperform poorly on field sobriety tests. It is, therefore, the intent of this study todetermine the effectiveness of a battery of field sobriety tests in aiding the officer todifferentiate between sober and intoxicated boat operators who have been exposed tothe same marine environmental conditions.

In practice, when an officer suspects that a boater is operating while intoxicated, he orshe will stop the boat and observe the appearance, demeanor, and behavior of thesuspect. The officer may then ask the suspect to perform certain field sobriety tests.The results of these observations and tests are used to determine whether to detain theboater in order to obtain the breath, blood or urine sample required to perform achemical test to determine the boater's BAC (Blood Alcohol Concentration). Deten-tion of an individual who does not exceed the legal limit is an inconvenience to theboater and a waste of time for the officer. This is true even in cases where the officercarries a portable breath tester because he or she must wait approximately ten minutesbefore taking the sample in order to eliminate the possibility that alcohol remaining inthe mouth will contaminate the sample.

When chemical tests of BAC are used as evidence, they are frequently challenged basedon the accuracy of the test instruments, the procedures followed, and the custody of theevidence. The results of valid field sobriety tests could be used not only to increase theaccuracy of arrest/release decisions made by the marine officer but also as evidence incourt procedures.

1.2 Study Background

Boating safety statistics compiled by the U.S. Coast Guard (Boating Statistics 1986)indicate that 1,066 lives were lost in 1986 as a result of recreational boating accidents.The role of alcohol as contributing to the cause of these accidents is not known.

In order to investigate the role of alcohol in recreational boating safety, the U.S. CoastGuard contracted with the Transportation Research Board of the National ResearchCouncil to identify and develop a list of research efforts which would improve ourunderstanding of the role of alcohol intoxication in marine safety and support lawenforcement efforts.

In February 1986 tlhe TRB report "Workshop on Alcohol-Related Accidents in Recrea-tional Boating" was published. The Coast Guard selected three of the efforts from thereport and arranged for support from the Transportation Systems Center in theirconduct.

The efforts selected were:

1. Assessment of the increased risk associated with alcohol intoxication and fatalaccidents in recreational boating.

2. Identification and evaluation of remote detection cues for alcohol intoxication inrecreational boat operators.

3. Assessment of the effectiveness of non-chemical tests of intoxication in the marineenvironment.

This report describes the third effort.

1.3 Background - Field Sobriety Testing

The police officer in the field may use a number of tools in assessing the sobriety of asuspect, including: observation, interrogation, a series of performance tests known as"field sobriety tests", and Portable Breath Testing (PBT) units. The evidence derivedthrough the accumulation of i.iformation provided by these indicators of intoxicationis used in the process of developing probable cause for an OUI (Operating Under theInfluence) arrest. One might ask, "If the officer can determine the suspect's BACthrough use of a PBT, why is it necessary to also give field sobriety tests?". PBTs doprovide a quick and reliable estimate of blood alcohol concentration (BAC). However,it must be understood that very few states have marine per se laws (laws which specify

2

the legal BAC level for intoxication) and implied consent laws (laws which requiiL aboat operator to submit to a PBT or other chemical BAC test). Even if a state has amarine per se law (OUI is defined as a BAC exceeding some number) and an impliedconsent law requiring a suspect to submit to a chemical test to determine BAC, a PBTreading alone is not necessarily sufficient evidence to support an O[I charge. Thesuspect may claim that the PBT unit was not properly calibrated, or that the officermisread the reading. Field sobriety tests provide evidence of impaired behavior whichcan rsupport the OUI charge. In general, the determination of a case is lSlally basedon the accumulation of evidence and rarely rests on only one indicator of possible (uilt.In this study the non-chemical behavioral indicators of intoxication are evaluated.

Currently, highway officers use a recognized battery of field sobriety tests, w, hich hasbeen validated as an effective means for assessing impaired behavior. A specifiedcriteria has been established for "normal" behavior within each of these tests. After theofficer has ruled out the possibility of impairment due to age, physical condition, illness,disability or fatigue, it is assumed that deviations from "normal" performance areattributable to intoxication.

The results of this testing are used to support the officer's contention that there existsprobable cause to arrest the operator. Many of these tests are also routinely used bymarine law enforcement officers in identifying the OUI boater. However, until nowthese tests have not been systematically examined for their validity of use in the marineenvironment.

An assessment of "Psychophysical Tests for DWI (Driving While Intoxicated) Arrest"was performed for the U.S. Department of Transportation, NHTSA, by Burns andMoskowitz (1977). They examined the effectiveness of six tests for their sensitivity aspredictors of impairment attributable to alcohol consumption. All six tests were foundto be "alcohol sensitive". The officers were found to make correct arrest/releasedecisions for 76% of the participants. The six tests studied were: One-Leg Stand, Walkand Turn, Horizontal Gaze Nystagmus, Finger to Nose, Finger Count, and Tracing. Ofthese six tests, Walk and Turn, One-Leg Stand, and Horizontal Gaze Nystagmus wereidentified as the three "best predictors" of intoxication and were recommended for useas an "abbreviated battery".

Tharp, Burns, and Moskowitz (1981) studied the effectiveness of the abbreviated testbattery. When officers used the Walk and Turn, One-Leg Stand, and Horizontal GazeNystagmus they were able to correctly classify 81% of the subjects as being above orbelow 0.10% BAC.

3

III tile present StidV. 1tie of tile Si'-, abov mtioned "alcohol senismvC" tests \rCeaHllie In the mlarinle ell% mioflieit. Tile Iraciu,, task was omitted Since Iaper and

l~icltak are1 dihciwt l oiters to p~erform (reizardles.s otlalcohol Conlllll)lio)

(ILIC 10 theC m(tion l af b oat aS it IS affected(I lv waves amid wakes. 'kvo additional fields bhrie tv et \ c re added to this basic I atte ry. si ice I hey are cuirrentI lvSC Ised in marinela1\ cmi twce menmt and are co nside red to be effective by hle stts uISI~ nt ~ ii ((CA.. NIDI.,anld (0)I .). !'liv are thle Alphabet Recital and I land Pat 'L St. These tests., are easily\anld (qu ckly :iiinistered. Ill the mar11ine eniviromniemit.

4

2 METHOD

This was a controlled experimental study in which a field sobriety test battery consistingof tFSTs commonly used in highway law enforcement were studied in the marineenvironment. Their effectiveness as predictors of BAC was assessed. Ninety-sevensubjects from a population of Coast Guard, Army, and Marine personnel in theYorktown, VA. area, participated in the study.

All study procedures involving the use of human subjects were reviewed by an "Insti-tutional Review Board", which was convened by Dunlap and Associates. This was anindependent committee whose primary concern was for the safety of the subjects. Allstudy procedures were approved by the committee.

Subjects were dosed, with measured mixtures of grain alcohol and fruit juice, to one ofthree levels of Blood Alcohol Concentration (BAC) in accordance with their drinkinghistory and body weight. The three target BAC levels were 0.00%, 0.08%. and 0.12%.Dosing occurred mainly on land (three drinks during a two hour period), with the final(fourth) drink consumed on the boat. Subjects were exposed to the marine environ-ment, i.e., a 90 minute boat ride as passenger. Subjects were passengers rather thanoperators for their own safety. There was no reason to allow the subjects to operatethe boats, since the study deals with the effects of exposure to the marine environmenton field sobriety test performance and not on boat handling.

A BAC measurement was taken on the boat prior to commencing field sobriety testing.The researcher measured the subject's BAC through use of a PBT The results werenot revealed to the subject or officers. Field sobriety tests were then given on the boatby a team of three marine law enforcement officers. One officer (rater # 1) served asthe lead officer, administering the tests while the other two officers observed. Allofficers administered the Horizontal Gaze Nystagmus iidividually, since it requiresface-to-face contact with the suspect in order to rate performance. After giving a testor pair of tests (as designated) each officer on the team gave a written estimate of thesubject's BAC level. Estimates of each officer were kept confidential so that one officercould not be influenced by the estimate of others. The subject was then transported toland. Field sobriety tests were then given on land by the same team of three officers.One officer (rater #2) served as the lead officer, administering the tests while the othertwo officers observed. All officers administered the Horizontal Gaze Nystagmus(HGN) individually. After giving a tes' or pair of tests (as designated) each officer onthe team gave a written estimate of the subject's BAC level.

The sequence of testing was similar to actual arrest procedures. In such proceduresthe officer begins by interviewing the suspect, proceeds to easily administered perfor-

5

mance tests, then administers the horizontal gaze nystagmus (provided that the officeris trained in administering this test). If at this point the officer feels relatively certainthat the per,;on may be intoxicated, he or she is transported to shore where balancetests can be administered.

Throughout the process the officer assesses the suspect's abilities and impairments andrevises his or her estimate of the suspect's BAC. The intention was to replicate thisprocess in this study.

2.1 Design

In order to determine the effectiveness of the field sobriety test battery in aiding theofficers to identify subjects who are intoxicated (generally at or above 0.10% in moststates), three experimental groups were dosed to reach target BAC levels and tested.Each of the thi ee groups was composed of approximately one third of the 97 subjects.

Subjects Dse Range Target BAC LevelGropA 32 0.10% to 0.14% 0.12%Group__B 33 0.06% to 0.10% 0.08%GroupC 32 0.00% 0.00%

In the original proposal for this study it was intended that the three groups would befurther divided in half in order to test for an order-effect for the tests believed to bethe least accurate (the Alphabet Recital, Thumb to Finger Count, Hand Pat, and Fingerto Nose). The orders of testing for each half of the subjects was to vary slightly, i.e., theorder of the four performance tests would be reversed and the order of the two balancetests would be reversed as follows:

ORDER 1 (49 Subjects) ORDER 2 (48 Subjects)

Interrogation InterrogationBehavioral Observation Behavioral Observation1st BAC Estimate 1st BAC Estimate

Alphabet Recital Thumb to Finger CountHand Pat Finger to Nose2nd BAC Estimate 2nd BAC Estimate

Finger to Nose Hand Pat

Thumb to Finger Count Alphabet Recital

6

3rd BAC Estimate 3rd 13AC Estimate

Horizontal Gaze Nystagmus Horizontal Gaze Nvstagmus4th BAC Estimate 4th BAC Estimate

On Land - flGN On Land - HGN5th BAC Estimate 5th BAC Estimate

On Land - Walk and Turn On Land - One Leg StandOne Leg Stand Walk and Turn

6th BAC Estimate 6th BAC Estimate

lowever, difficulties arose in obtaining a firm commitment for the participation of all96 subjects. The experimental design used required 96 subjects. When the study beganit was uncertain whether enough subjects would participate and the decision was madeto begin testing subjects using Order I only and abandon the effort to test for an ordereffect.

2.2. Field Sobriety Tests Used

Interrogation and Observation

During a routine OUI investigation, the initial contact which the officer has with thesuspect provides a period for interrogation and observation. During the first fewmoments of contact, the officer engages the suspect in conversation in order to havean opportunity to observe the suspects ability to answer simple questions and todemonstrate orientation to person, time, and place.

While conducting the interrogation the officer makes observations regarding thesubject's appearance and manner. Observations may include cues obtained throughchecking: clothes, breath, attitude, facial coloration, eyes, pupils, speech, unusualactions. Often officers in the field are not provided with a formal checklist of items,but rely on recall in making their observations.

In order to insure that both teams of officers conducted a similar observation process,a list of typical questions was provided from which the officer could conduct theinterrogation process. Officers were instructed to phrase the questions in their ownstyle so that language, and the situation in general, would not be awkward or stilted.

This list was extracted from the Ohio Department of Natural Resources, Division ofWatercraft, "Alcohol Influence Report". Similar checklists are in use in California,

7

Maryland, and in several other states. Appendix, item 1 includes a copy of interrogationquestions, observation checklist, and scoring, sheets for all testing.

Alphabet Recital

The suspect is asked to recite the alphabet from A through Z. In some locales, officersask the suspect to recite from the middle of the alphabet, specifying a letter to beginat, for example, "recite the alphabet starting from the letter Y. This is done. since itrequires more thinking on tne part of the suspect and may show confusion and lack ofreasoning on the part of the suspect. However, for purposes of the study, the suspectwas asked to recite from A through Z. In some court proceedings judges have beenknown to consider mid-alphabet recital as an attempt by the officer to confuse thesuspect. Therefore, the method most widely accepted in a court of law as evidence ofpossible intoxication was used in this study.

Hand Pat

The suspect is instructed to hold both palms out, facing up. The left hand is keptstationary while clapping the palms together. When the palm is struck the personcounts ONE. He then turns the right hand over and claps the back side of the palmand counts TWO. The suspect continues to clap alternating palm and back side of palmand counting ONE, TWO. The suspect is asked to count out loud and to increase hisspeed clapping and counting.

Finger to Nose

The suspect is seated and asked to put his hands at his side. With eyes closed and headtilted back slightly, he is asked to touch the tip of his nose with the tip of his index finger.When the officer says RIGHT the person uses his right hand. When the officer saysLEFT the suspect uses his left hand.

Finger Count

The suspect is asked to touch and count each finger in succession, counting aloud. Hetouches thumb to finger and counts the four fingers aloud 1-2-3-4 and then reversescounting 4-3-2-1. He is instructed that each time he counts he should try to go a littlefaster.

8

Horizontal Gaze Nystagmus (I-GN)

This test measures the involuntary lateral jerking motion of the eyes. It occurs uponlateral gaze when BAC exceeds .06 (Burns and Moskowitz, 1977).

The suspect is asked to look at the tip of a pencil. Keeping his head still, he is asked totrack the tip of the pencil with his eyes while the officer moves it.

Walk and Turn

This is a test of balance and was, therefore, given on land. It is also a test of the person'sability to follow simple instructions and to divide his attention between listening to theofficer and maintaining a specific standing position.

The suspect is instructed to stand on a line with his right foot in front of the left. He isinstructed that the right heel should touch his left toe.

The suspect is instructed to take NINE heel-to-toe steps down the line, turn around,and take NINE heel-to-toe steps back. In turning around, the suspect turns by pivotingon one foot. He is asked to keep the foot on the line and use the other foot to turnhimself around with several small steps. Hands are kept at the sides at all times. Thesuspect is instructed to watch his feet at all times, and count the steps out loud.

One-Leg Stand

This is a test of balance and was, therefore, given on land. It is also a test of the suspect'sability to follow simple instructions.

The suspect is asked to stand with his heels together and his arms down at his sides. Heis asked to raise one leg so that his heel is about six inches off the ground and to holdthat position. While watching his raised foot the person is to count from 1001 to 1030.

2.3 Subjects and Raters

2.3.1. Subject Description

Ninety-seven subjects participated in the study. The subjects were all military person-nel from various facilities in the Yorktown, VA. area. The majority of subjects wereCoast Guard personnel from The Reserve Training Center (RTC) Yorktown, the siteof the study. The remainder of subjects were from Coast Guard Fifth District, Marinesfrom the Naval Weapons Station, and Army personnel from Fort Eustis.

9

Subjects participated on a voluntary basis. They were given a number to be Usedthroughout the study and assured that all data would be confidential. Men between theages of 2 1 to 42 were the participants. The rationale for concentrating testing on thenmale rather than a mixed male-female sample is explained below. The appendix, item2 lists the age distribution for participants.

Range = 1 to 42 yearsMean = 27 yearsMode = 2 1 years (21 (7 Of all subjects)

In a review of arrest data for intoxicated boat operation for the two year period of 1985and 1986 it was found that the over-whelrning majority of arrests involved men withinthe ge group of 2 1 to 40. This review of arrest records was initially performed forlThsk

2 of this project, w hich involved the identification of possible remote detection cues ofintoxication. Arrest records were reviewed in the two states of Ohio and Maryland andtwo counties in California (San Joaquin County and Lake Shasta). Furthermore, inTask I of this project, Accident Reports of fatal boating accidents occurring in Califor-nia and North Carolina (the two states which have kept the most complete records offatal boating accidents) were reviewed. This data indicated that the majority of fatalboating accident victims were men.

2.3.2. Subject Screening

Screening was conducted by Dunlap and Associates. A notice was posted or appearedin the newsletter of the facilities from which subjects volunteered. Potential subjectswere informed that this was a controlled study concerning the effects of alcohol onboaters. Male subjects between the ages of 21 to 50 were solicited. A sample an-nouncement, which was printed in "Plan of the Day" (RTC Yorktown newsletter),appears in appendix, item 3.

Potential subjects were initially screened either on the telephone prior to participationor in-person immediately before participation. They were asked about their age,boating experience, susceptibility to seasickness, medical condition, and drinkinghistory. The "Telephone Screening Instrument" appears in appendix, item 4. Based onanswers to these questions, the person was selected or rejected for participation. Ifselected, an appointment was set for participation in the study.

During a face-to-face interview (either prior to the day of participation or immediatelybefore participation) the subject completed an "Alcohol Questionnaire" (see appendix,

10

item 5). The results of this questionnaire were used by Dunlap and Associates inassigning subjects to one of the three dosing groups. This was done in order toappropriately assign drinkers in accordance with their drinking history and habits.

This instrument produces two scores. The first score is the sum of the response weightsfor Question I through 11, and reflects the quantity, frequency and circumstances ofthe subject's typical drinking situations. The second score is the sum of the responserates for Questions 12 through 19, and reflects the subject's manifestation of generallyaccepted indications of "heavy" drinking. Questions 12 through 19 were derived fromthe Michigan 'Alcoholism Screening Test".

Previous applications of the instrument by Dunlap and Associates led to establishmentof the following "heavy" drinker score threshold:

1. a score of 25 or more on Questions 1-11, irrespective of the score on Questions 12-19;or,

2. a score of 18-24 on Questions 1-11, provided that a positive score (1 or more) isobtained on Questions 12-19.

The questionnaire score criteria used for the assignment of subjects to the three groupsis included in appendix, item 5.

2.3.3. Raters/Marine Law Enforcement Officers

Two marine law enforcement officers from Maryland Department of Natural Resour-ces and four marine law enforcement officers from Ohio Department of Watercraftserved as raters for the study. All officers were experienced in field sobriety testingand arrest procedures.

Officer experience in marine law enforcement ranged from four to fifteen years. Themajority of officers had nine or more years of experience (only one officer had onlyfour years of experience). All the officers had completed extensive training in fieldsobriety testing. They were experienced in the use of field sobriety testing and hadspecific certification in Horizontal Gaze Nystagmus. Five of the six officers workedprimarily in the field, while one of the officers worked primarily in an administrativeand training capacity. Specifics on each of the officers are listed in appendix, item 6,"Years of Experience in Marine Law Enforcement and OUI Arrest Experience ofOfficers".

11

2.4. Facilities and Equipment

Facilitics and equipmenit were jointly provided by RIC Yorktown and TSC. RTCYorktown provided tihe physical site for the study incluiding the rooms used for subjectpreparation and recoveir. TSC provided all breath testing equipment.

2.4.1. Study Site

I-wo rooms in the GVmnasiu m Complex at RTC Yorktown were used for briefing,dosing, and recovery of subjects. Room #1 was used for the medical screening ofsubjects which took place prior to dosing. It also housed the breath testing equipmentand breath-technician, dosing apparatus and dosing-technician. Room #2 was the areain which the subjects drank their drinks. Cards, magazines, and movies were providedfor their entertainment during dosin. This room was also used for recovery, i.e., afterbeing tested subjects returned to this room while waiting for their BAC to be confirmedto be ().0(Q prior to release.

Wormley Creek was the docking area from which the boats were launched and re-turned. Figure 1 depicts the waterfront area. This area is about half a mile from theGymnasium Complex. Therefore, a van was used to transport subjects to the dock.

2.4.2. Boats

Motor boats in the 16 to 18 foot range were used to provide the 90 minute boat ride foreach subject. These were boats rented from U.S. Army, Fort Eustis. This particulartype and size of boat was used since "Boating Statistics 1986" compiled by the U.S. CoastGuard indicates that the majority of fatal recreational boating accidents occur in boatsthat are less than 26 feet long. In addition, the previously mentioned arrest dataindicated the majority of OUI arrests to involve operators of boats in this type and size.

The boats used by the officers from which they conducted the on-water testing were 16to 18 foot "Boston Whaler" type boats. They are typical of the boats they would usuallybe operating on patrol.

2.4.3. Dosing Schedule and Apparatus

Appendix, item 7, "Dosing Tasks" outlines the schedule followed for dosing subjectson each day. Some variations from the schedule occurred due to no-shows and latearrivals of subjects.

12

Figure 1. The Waterfront Area

Appendix, item 8 lists the "Nominal Dosing Levels" used for the subjects. Dosing wasdone with 190-proof (95%) grain alcohol. The drinks consisted of the appropriatealcohol dose and approximately 9 ounces of orange, grapefruit or tomato juice. Thekind of juice used was dependent on the preference of the subject. For the Group Csubjects (0.00% BAC), the drinks consisted of the preferred juice with approximately4 milliliters of grain alcohol "floated" on top (i.e., the Group C subjects consumed atotal of about 16 milliliters of grain alcohol over the four drinks served.) In no case didthe Group C subjects register other than 0.00% on any of the breath tests given. Inaddition, no subject registered other than 0.00% on the first breath test of the day.

The actual doses given to the subjects varied somewhat from the nominal figures. Earlyon, the attempt was to "fine-tune" for the on the water testing. There were alsovariations in individual absorption rates due primarily to differences in stomach con-

tents at the start of the dosing sessions. Although subjects were asked to eat only a"light" breakfast before arriving for participation in the study, some subjects did notcomply. Dose levels were adjusted to attempt to compensate for such factors.

13

The actual dose received by each subject along with the subjects number, groupassignment, and body weight are presented in appendix, item 9.

2.4.4. Breath Testing Equipment

Breath testing equipment was provided by TSC. All breath testing done on land, i.e.,in the dosing area, was performed on the Seimens Alcomat. This highly accurateinstrument uses an infrared absorption sensor. It provides a three-digit read-out ofBAC. The machine provides a digital display on a small screen and also a printeddisplay on a paper tape. For all on land breath testing there were two tests performedfor each round of testing in order to insure the accuracy of the reading.

Breath testing done on the boat, prior to the officers beginning the field sobriety testing,was performed with the Alcometer S-D2, a product of Lion Laboratories, Ltd. Thissmall, portable, hand-held, light-weight instrument uses a fuel cell sensor. It providesa modified three-digit display of BAC. The last digit is rounded to the closest 0 or 5reading, for example, 0.054% is rounded to a reading of 0.055%.

In order to determine the accuracy of the rater's estimates of BAC these estimates werecompared to the BAC readings taken with the Alcometer PBT prior to field sobrietytesting. For purposes of discussion the PBT readings are referred to as the "actualBAG" even though they are really breath based estimates of arterial blood alcoholconcentration.

The accuracy of the reading obtained from the fuel cell PBT reading was checkedagainst the Gas Chromatograph. Each PBT reading was compared to the readingsobtained through breath testing using the Seimens Alcomat. Each PBT reading wascompared to the Alcomat reading taken prior to boarding the boat and the readingtaken immediately upon return to the recovery area following field sobriety testing. In96 of the 97 cases the readings from both instruments (PBT and Alcomat) werecompatible. One reading on the PBT was unrealistically high, which may have beendue to the subject belching and, therefore, elevating the PBT reading. In this one casethe actual BAC was adjusted for the analysis, i.e., made consistent with the Alcomatreadings.

Both types of instruments were used by trained study personnel and the BAC readingswere not revealed to the officers or subjects throughout the testing day. Both types ofinstruments were calibrated at the begining of the experiment and after the first fivedays. Both instruments were used with a disposable mouth-piece for each subject.

14

2.4.5. Routine and Emergency Medical Personnel

Medical personnel were provided by RTC Yorktown in order to conduct a routinemedical examination prior to each subject's participation in the study. Appendix, item10 is a copy of the "Medical Report" form.

During the examination the following items were checked: pulse, blood pressure,temperature, respiration. The subject was also questioned regarding recent consump-tion of any prescribed or over-the-counter medication. Based on the results of thisexamination, the person would be certified as qualified or not qualified to participatein a controlled drinking study.

Throughout the study emergency medical assistance was on-call. RTC Yorktown hason-grounds medical facilities which were available for use in the event of illness orinjury of any participant.

2.5. Preliminary Procedures

This section describes in detail the preliminary procedures followed before actualtesting. These procedures include the establishment of a testing schedule, prerequisiteenvironmental conditions under which testing could occur, instruction to subjects andofficers regarding participation.

2.5.1. Testing Schedule

Two subjects were scheduled to arrive per hour. The first two subjects arrived at 07:00hours. With this schedule up to fourteen subjects per day could be tested duringdaylight hours. The target number for subjects was twelve per day. Due to no-showsand cancellations, on several days less than twelve subjects were tested. Subsequently,on other days extra subjects were tested. The most subjects tested in one day wasthirteen. All testing was completed within nine days.

2.5.2. Environmental Conditions

All testing was conducted from May 11 through May 21, 1987 on the James River inYorktown, VA. Testing was conducted during relatively mild summer-like weatherconditions. The intention was to conduct testing during the typical weather conditionsexperienced by the average recreational boater. Furthermore, "Coast Guard BoatingStatistics" indicate that the majority of fatal boating accidents occur during daylighthours in calm waters, little or no wind, good visibility, and water temperature of 60 to79 degrees fahrenheit.

15

No tCstinL' vas conducted Under adverse conditions, such as, hi i70h winds and disturbedsca state. ( )n the average the temperature was approximately 80 to 85 degrees, rangi figfrom 05 to 95 decrees. "lesting was conducted on both sunny and cloudy days. Testing%vas also conducted under conditions of light rain. No testing was conductinrig duringheaivy rain. All testilng was conducted during daylight hours.

2.5.3. Subject Briefing

Subjects were informed regarding the day's proceedings as soon as they arrived forparticipation ("Briefing Sheet", appendix, item 11). During the briefing subjects wereasked to sign an "Informed Consent Form" (appendix, itern 12). They were told thattheir participation would be for approximately seven or eight hours. They were toldthat "I mledical technician would be giving them a brief mcdical examiilation in orderto certify that they were ready for participation.

Each subject was told that he would be expected to consume a glass of fruit juice witha measured amount of alcohol each 40 minutes until three glasses had been consumed.Before and after each drink he would be given a breath test. Following the third drinkand breath test, he would be asked to board a boat for a 90 minute boat trip. Duringthe trip he would be asked to consume a fourth drink. After the boat trip a breath testwould be taken on the water and the subject's boat brought along side the officers' boat.The three police officers would then give the subjects standard field sobriety tests.

Subjects were instructed to act as if the situation was an actual encounter with a policeofficer and as if the person was indeed in danger of arrest. Subjects were instructed toperform as well as possible on each test. The seriousness of the scenario was em-phasized. It was the job of the subject to convince the officers that he was sober.

Subjects were also informed that after testing they would receive a meal and periodicbreath tests in order to monitor the elimination of alcohol from their bodies. Whenthe alcohol content level was at 0.00% the subject would be released.

2.5.4. Officer/Rater Briefing

Prior to beginning on Day 1 the officers were trained in the use of the "Scoring Sheets".It was explained that the six officers would be working in two teams of three officers.Offi,:ers were designated as belonging to Team 1 (Rater # 1, #2, #3) and Team 2 (Rater# 1, #2, #3). Two subjects would be tested simultaneously, i.e., one being tested byeach team.

16

Rater # I from each team served as the lead-officer during the on-the-boat testi ng.Rater # I would greet and question the suspect and give all instructions for TheAlphabet Recital, lI anid Pat, Finger to Nose, Finger Count. All three officers w\(ouldbe observing for the behavioral indicators of intoxication and w( ould be observing theresults of the tests given by the lead-officer. The I IGN was to he given by all threeofficers, since this test requires a face-to-face position by each suspect and officer i1Aorder fo the officer to make an accurate rating. Therefore, the lead-officer (Rater # 1)would give the I IGN, then Rater #2, and then Rater #3 would give the test.

For the on-the-land testing again each officer gave the HGN. Then Rater #2 from eachteam served as the lead-officer for the two on land tests: The Walk and Turn, andOne-leg Stand.

During the course of testing each officer was to make a BAC estimate after each testor pair of tests (as designated). With the results of each test the officer would havemore information on the subject's abilities or level of impairment. Therefore, the BACestimate mav change as more information is received. Officers were instructed to maketheir estimates in accordance with their opinion at the moment, based on what they hadseen thus far.

Officers were instructed to keep their scoring and BAC estimates confidential. Theywere not to look at each other's score sheets and were not to discuss their opinions onthe subject's performance or BAC level. They were also cautioned to be mindful oftheir gestures and facial expressions and to guard against indicating their opinions inthis manner. They were informed that an observer would be present at all timesthroughout testing in order to inform them of any cues which they may inadvertentlybe giving to each other.

2.6. Study Procedures

This section describes, in detail, the procedures experienced by the subjects and officersthroughout their participation in the study. Procedures are discussed for: preparingsubjects for dosing, dosing of sUbjects, exposing subjects to the marine environment,field sobriety testing on the boat and on land, and subject monitoring and release.

2.6.1 Subject Arrival, Briefing, and Medical Examination

As previously mentioned, subjects arrived in groups of two. Two subjects arrived eachhour beginning at 07:00 through 12:00. Subjects were briefed regarding their participa-tion in the study. Subjects completed the screening questionnaires (if not previouslycompleted prior to the day of participation). The medical examination was given. Each

17

subject was then certified a, ready or iot ready f(br participation by the medical

Ihc dosing technician tueLStionCd the stIblcct regardinu what he had for breakfast inr)FdCIe to adjLIst the dose coilsideringt differences in absorption rates due to stonach

c mtent. Sutj ecis had been asked to limit their breakfast to a "liht meal" of toast orcerea, I\\ er, not all subJects adhered to the request awl it was necessary totiLI sIMo them re ,zrcliing their meal.

Ilic d(ioi17 lechiicia i then iixed the drink in order to begin elevati(i or no elevationOo)r he placbo grou)) O1 the subject's BA level in accordance ith the subject'sOr(I' ii .,,swiiinicnt. Beftore giving the ,il)'ect the first crink a breath sample was takenh\ the hreath tuchnician.

I i p c'ess L(ook approxinateki 3(1 liLtites. F(or a sLibject arrivinr at 07(0(0. lie woUldbe read\ for dosing, at 0)7:30.

2.6.2. Dosing on Land

The first drink was consumed during the first 20 to 3) minutes of the 40 minLIte period,then iafter vaiting tell Minutes a breath test was :. T!hr,)ughout the study, thestibect, did not drink or smoke for ten miii , o iU, i) reath testing, since this wouldhave interfered w\'h a proper BA( aoing. At no time prior to field sobriety testingdid sLIbjectS conIslmle food, since this would have affected the subject's alcohol absorp-tinn rate. Movies, magazines, aid c ,.: ,. ,c 2-mided for the subjects to entertainthemselves during dosing

Drink #2 was given and consumed during the first 20 to 30 minutes of the next fortyminute period. After waiting ten minutes a breath test was given. Drink #3 was givenand consimned during tile first 20 to 30 minutes of the next forty minute period. Afterwaiting ten minutes a breath test was given.

Following the consumlption of the three drinks, each subject was escorted to a boat. Anescort acconipanied tile subject in order to insure his safety. The launching area wasapproximately a quarter-mile from the dosing area. The subjects were transported byvan to Wormley Creek where they boarded a boat. Generally one subject rode on theboat with one boat operator and the escort. Due to periodic mechanical difficulty withsome of the boats, at times two subjects rode in the same boat.

18

2.6.3. Dosing on the Boat

A thtie Sh*t 1ha 1 been~ ra for approximnately tell 11iirn teS hie was orven the

fourth drink. '['his drink was pre-nilixed hytedsng technician on shore and carriedalon', by the esco)rt.

2.6.4. Rendezvous wvith the Police Boat

\fte r K)*1( i 11 rfuncra or iii ie tv mn U te5. tilie boa t operators radioed thle lhoat laUnrCha rca to an n1ounrce tie ir aIpp roach. L~acli teamn oh office rs were wait i n aboard a boat forone subl)Ject to test. OnIe subiCt boat and one police boat were rafted together. (ThietV(sets of',u bject arid 1)0C ice oats were stat iorid approx i matelv I doto 200yards from!feaich other, i.e., a1 sulicierit distance to insure non-interference.) The officers uIsuallystayed intheir boats (three in each boat) while the subhjects staved in thieir boat (one ineach boat, accoripariied hyesco rt anid operator). Ini the field ol 'icers uIsully11 condLIctfield sobriety testing fromn their own boats and do not board the suIspect's boat Unlessnecessary. The officer is at risk when lie boards the su~spect's boat. Figure 2 shows theposrtion of boats dUring testing.

Figure 2. Position of Boats During Testing

19

2.6.5. Testing BAC on Boat

Before the officers began die interroLation. observation. and field sobriety testin, aBAC test was civen to the subject by the researcher. The test was performed with theportable breadth unit and the test results \were kept secret from the subject and officers.Figure , depicts the breath testing. This test was perforned in order to determine thesubject's BAC level prior to testing so that, during data analysis, the officer's estimatedBACs could be compared to the actual BAC. Figure 4 shows the BAC levels of thesubjects measured on the boat immediately prior to FST The BACs measured arefound in Table I.

2.6.6. Field Sobriet\ Testing on the Boat

The procedure described to the officers during the 'Officer/Rater Briefing" \wasfollo\ed during the nine day testing period. Testing occurred as follows:

The lead officer (who was Rater #1) designated to perform the on boat testing beganquestioning the subject, using the "Interview" section of the scoring sheet as a model.

Figure 3. Breath Testing Prior to Field Sobriety Testing

20

170

Q0

o ____ wL0~C0

o Co

Q0 cc

21\

l'Abl t

BACs MEASURED PRIOR TO FSTSubject # * % BAC Subject # * % BAC

Subject 1 .00% Subject 49 07%Subject 2 .00% Subject 50 .07%Subject 3 .00% Subject 51 .07%Subject 4 .0000 Subject 52 .07%Subject 5 .00% Subject 53 .07%Subject 6 .00% Subject 54 .07%Subject 7 .00% Subject 55 .07%Subject 8 .00% Subject 56 .07%Subject 9 .00% Subject 57 .08%Subject 10 .00% Subject 58 .08%Subject 11 .00% Subject 59 .08%Subject 12 .00% Subject 60 .08%Subject 13 .00% Subject 61 .08%Subject 14 .00% Subject 62 .08%Subject 15 .00% Subject 63 .08%Subject 16 .00% [Subject 64 .09%Subject 17 .00% [Subject 65 .09%Subject 18 .00% [Subject 66 .09%Subject 19 .00% [Subject 67 .09%Subject 20 .00% [Subject 68 .09%Subject 21 .00% [Subject 69 .10%Subject 22 .00% [Subject 70 .11%Subject 23 .00% [Subject 71 .11%Subject 24 .00% [Subject 72 .11%Subject 25 .00% [Subject 73 .11%Subject 26 .00% iSubject 74 .11%Subject 27 .00% [Subject 75 .11%Subject 28 .00% [Subject 76 .11%Subject 29 .00% ISubject 77 .11%Subject 30 .00% ISubject 78 .11%Subject 31 .00% ISubject 79 .11%Subject 32 .00% ISubject 80 .11%Subject 33 .04% [Subject 81 .12%Subject 34 .04% [Subject 82 .12%Subject 35 .04% Subject 83 .12%Subject 36 .05% [Subject 84 .12%Subject 37 .05% Subject 85 .12%Subject 38 .05% [Subject 86 .12%Subject 39 .05% [Subject 87 .12%Subject 40 .05% ISubject 88 .12%Subject 41 .06% Subject 89 .13%Subject 42 .06% Subject 90 .13%Subject 43 .06% Subject 91 .13%Subject 44 .06% Subject 92 .13%Subject 45 .06% [Subject 93 .14%Subject 46 .06% [Subject 94 .15%Subject 47 .06% [Subject 95 .15%Subject 48 .06% [Subject 96 .15%

Subject 97 .16%

* Subject numbers do not represent order of dosage

22

The offcer phrised the questions in hils own words. \Ieanwhile. the other two officerswere listening and watching. All three officers made observations in accordance With

the "()hservation" section of the scorno sheet. Then each of the three off'icers made aBAC estimate on his or tier scoring sheet. Each officer could not see the estilate ofthe other officers. Then Rater # I began the field sobriety testing. The "AlphabetRecital" was given IK Rater # I and individually scored by each rater in confidence onhis/her scoring sheet. The " land Pat" test was given Ky Rater # I and individuallyscored Ky each officer in confidence on his or her scoring sheet. After these two test,had been ven each rater wrote down a BAC estimate for the subject. This estimatewas based on the information received thus far.

Then Rater # I gave the subject the "Finger to Nose" test. Each rater individually scoredthe subject's performance. Rater # 1 then gave the subject the "Liinger Count" test andagain each rater individually scored the subject's performance. After these two testhad been given each rater wrote down a BAC estimate for the subject. This estimatewas based on the information received thus far.

Rater # I gave the HGN test (depicted in Figure 5) and rated performance and wrotedown a BAC estimate. Then the other two officers in turn gave the HGN, ratedperformance, and wrote down a BAC estimate. Once again, this estimate was basedon the information received thus far.

Figure 5. Administering the HGN Test on the Boat

23

At this point the subject was transported back to land. Testing continued on land aftera brief hiatus, established to simulate the time which would be required to bring asuspect to shore in a real arrest. This wait did not significantly effect the subjects BAC.

2.6.7. Field Sobriety Testing on Land

Rater #2 in the team of three always served as the lead officer on land. This was donein order to reduce variance associated with subject exposure to different lead officers.

The icad officer on land gave the HGN, rated performance and gave a BAC estimate.Then the other two officers in turn did the same.

The subject was given ten minutes to attain "land-legs", i.e., stabilize balance in orderto perform the balance tests to follow.

The lead officer gave the "Walk and Turn" test(depicted in Figure 6) while the other two rat-ers observed. All three officers individuallyrecorded performance and wrote down a BACestimate. This estimate was based on the infor-mation received thus far. The lead officer gavethe "One-Leg Stand" test while the other twoofficers observed. All three officers individu-ally recorded performance and wrote down aBAC estimate. This estimate was based on theinformation received thus far.

2.6.8. Monitor and Release of Subject

The subject was then transported by van backto the Recovery Area. A breath test was takenon the Seimens Alcomat in order to later con-firm the reliability of the on boat BAC reading Jtaken with the portable device. The subject wasthen given lunch. Breath tests were given peri-odically and when the subject reached a 0.00%BAC he was released. igure 6. Administering the Walk and

Furn Test on Land

24

2.6.9. Officer Debriefing and Feedback Sessions

Each evening, following the testing of th last two subjects, a debriefing session washeld. Any logistical problems which had arisen during the day were discussed. Theofficers were told the actual BAC for each subject tested that day. This feedbackregarding actual BAC is representative of realistic arrest procedures. In the field, afterthe suspect has been given a chemical BAC test, the officer is told the score by thetesting technician.

25

3 ANALYSIS

As indicated in the -Methods" section, the raters estimated the subject's BAC imme-diately after administering the tests listed below. The "First" BAC estimate made bythe rater was based on the information obtained from "Interview and Observation".The "Second" BAC estimate was based on the infoimation from "Interview andObservation" and the additional information which may have been obtained from"'Alphabet Recital" and Hand Pat". Each successive BAC estimate was based on theinformation from all tests given to that point in time. Therefore, this was an accumu-lative process. As each test was given the assumption was that the rater would beobtaining more information on which to base the next BAC estimate. By the "Sixth"BAC estimate the rater's estimate was based on all information provided by adminis-tering all tests. The "Sixth" estimate is, therefore, referred to as the "Final CumulativcEstimate", i.e., resulting from the accumulation of information obtained from admin-istering all tests.

Tests Administered BAC EstimatesInterview and Observation FirstAlphabet Recital and Hand Pat SecondFinger to Nose and Finger Count ThirdHorizontal Gaze Nystagmus on Boat FourthHorizontal Gaze Nystagmus on Land FifthWalk and Turn, One Leg Stand Sixth (Final Cumulative Estimate)

Data were analyzed for all 97 subjects. As indicated above, there were six BACestimates made by each of the three officers assigned to evaluate the subject, with oneexception. Due to mechanical difficulties with one of the boats, the on boat HGN couldnot be given to one subject. Therefore, all analysis represents results for 97 subjectsfor the first, second, third, fifth, and six estimate, and 96 subjects for the fourth estimate.

All BAC estimates of all raters were included in the data analysis, with (:ne exception.One of the six raters (Team 1, Rater 3) was experienced in the administrative andtraining aspects of marine law enforcement, but did not have experience in theimplementation of OUI arrest procedures in the field. (Appendix, item 6 includes alist of the experience of all raters.) This particular rater had no recent history of makingarrests, while the other five officers had such OUI arrest experience. In examining thecorrelation between individual officer's final cumulative rating (BAC estimate givenafter administering all tests) and the BAC (as measured by PBT), it was found that thisofficer's first three days of performance in the study were very erratic. As the studyprogressed, performance for this officer improved dramatically. By Day 4 the correla-tion between the final cumulative rating and the measured BAC had stabilized.

26

Therefore the data analysis includes only those ratings provided by the officer after daythree. Improvement was probably achieved through exposure to the testing in the fieldand the daily Feedback Session described above.

The analyses performed for this report provide three different measures of perfor-mance:

* Correlation between the officers' estimates of BAC and the measured BAC's;* the absolute differences between the officers' estimates of BAC and the

measured BACs; ande the frequency with which the officers' estimates of BACs were correctly above

the BAC which would result in conviction (True Arrest) or correctly belowthe BAC which would result in conviction (True Release).

In order to satisfy the basic goal of the study: to determine if use of FSTs on individualsexposed to a marine environment would reduce FST accuracy, the performance of theraters in this study was compared to the performance of similar raters in studiesconducted in non-marine environments.

In prior non-marine studies of the effectiveness of FST's (Burns and Moskowitz,andTharp, et al.) Pearson correlation coefficients were computed between the measuredBACs and the BAC estimates of raters using FSTs. To determine if the accuracy of theraters in this study were consistent with prior studies, Correlation coefficients werecomputed between the measured BACs and the mean estimates of:

- both teams of raters, (the mean BAC estimates of the three teammembers for each of the two teams was compared with themeasured BAC of subjects each team rated)

- each rater individually (the BAC estimates given by individualraters were compared with the measured BACs of the subjectsrated)

Although the study's primary goal was to judge the relative accuracy of FSTs in marineand non-marine environments, the data gathered supported estimates of the absoluteprecision of the raters performance. This index of accuracy was the absolute deviationbetween the raters' BAC estimates and the measured BACs. This measure is moreeasily interpreted than the correlation and sensitive tests of the statistical significanceof differences between the accuracy of different portions of the FST battery areavailable (Student's t test).

Finally, in order to predict how well a trained marine safety officer could use the BACto discriminate legally intoxicated suspects from legally sober suspects "True Arrest"and "True Release" indices were developed by TSC. These indices describe the raters'

27

performances in determining whether a subject's BAC exceeded a set point. Thisprocedure provides information on the probability that an officer will correctly dis-criminate between individuals who are legally intoxicated and those that are not. These"True Arrest" and "True Release" indices are valuable for illustrative purposes but arenot parametric in nature and do not permit the use of sensitive statistical tests.

28

4.0 RESULTS

4.1 Correlations Between Estimates of Bac and Actual BAC

4.1.1 Overall Correlation Between Bac Estimates and Actual BAC

Figure 7 depicts the Pearson correlation coefficients (R's) calculated for all combinedraters, i.e., estimates on all 97 subjects (each correlation represents the data of allsubjects). It shows that in this study there was an increased correlation between theofficers' estimates and the actual BAC as the officers proceeded from the first estimatethrough the sixth (final cumulative estimate). As more information on the subject'slevel of impairment became apparent through administration of the tests, the officers'estimates more closely approximated the actual BAC.

While the methodology used by Burns and Moskowitz, (op.cit.) is not completelycomparable it is interesting to note that the mean correlation for all raters betweenmeasured BAC and the final cumulative rating was .704 and the value found by Burnsand Moskowitz, (op.cit.) was .669.

4.1.2 Correlation Between Individual Rater's Estimates and Actual BACs

As an aid in examining the performance of each rater and the ability of the field sobrietytest battery to improve the individual rater's estimates, Table 2 lists the correlationsbetween individual rater's estimates and actual BAC. These correlations are depictedgraphically in Figure 8.

4.2 Tests of Absolute Difference Between Estimated Bacs and Actual BACs

The absolute deviation between the BAC estimate and the actual BAC is expressed asa % BAC. A highly accurate rater's estimates on the average differed from the actualBACs by 0.023% BAC and a less accurate rater's estimate differed by 0.074% BAC.For example, a highly accurate rater's estimate might be 0.080% BAC when the actualBAC is 0.103%, (a difference of 0.023% between the estimate and the actual BAC).Paired t tests were used to determine if the differences of the absolute values of thedifferences between PBT measured BACs and officers' estimates of BAC at thedifferent stages in the FST procedure could have been attributed to chance factorsduring the data collection. This hypothesis is known as the null hypothesis. Thealternative hypothesis would imply that the differences observed in the study are likelyto be found in real world applications of the portions of the FSTs. A t value implies acorresponding probability (P) value.

29

A\N

-O m

300

(a I W ~ ~

U)

>1 0z

El): 4-3 ~- (N C--i U1(N

4J)

>4 0

41J

rs 00 0cl kooN Hr: Z U k \LfO 0 N (

4L f4 . . . . .

4--H

0

-HW w NmwZVc o

0

3c ., ~C C J cl ) Zrq -

4) 41 . . .)

4J

H -4EI -4 H N N

31

71-

C) ~ J u Ck- - 4 --

4-j U)

(4 Qo

NF :

CC) CQ P-

32N

This P value is the likelihood or probability that the observed results would occur if thenull hypothesis were true (the differences found between two phases of the FST couldhave occured through chance variation rather than some quality of the test). Todetermine if the accuracy of the estimates of BAC, averaged over all raters, differedsignificantly between the first estimate (absolute deviation = 0.047% BAC) which wasbased on interview and observation and the last estimate (absolute deviation = 0.027%BAG) which was based on the entire field sobriety test battery, a paired t test wasperformed. The results of the t test indicated that the difference in accuracy was,tatisticallv significant and had a probability of being due to chance of less than 0.001%(t = -4.19). Therefore, the raters' estimates of BAC using the FST were significantlyimproved over those achieved based on observation and intcrrogation.

In order to determine if there was an overall improvement in accuracy between the onboat nystagmus test and the on land cumulative estimate a t test was performed. Theaccuracy of the estimated BACs of all raters combined did not differ significantly fromthe fourth estimate 0.029% BAC (Gaze Nystagmus on boat) to the last estimate basedon the entire field sobriety test battery, including on-land testing 0.027% BAC (t = -

1.32). Thercfore, it would seem that on the whole, the small improvement in accuracybetween the estimate following the on boat HGN and the cumulative estimate was notstatistically significant. The group of officers did not improve the accuracy of theirestimates, as a result of the on-land testing.

Table 3 represents the mean absolute differences between estimated and measuredBAC for each rater for the first, fourth, and sixth estimates.

Table 3. Mean Absolute Differences Between Estimated and Measured Bac

Obs&InteLr OnBoatNys. Cure.Rater 1 TI 0.052% 0.034% 0.040%Rater 2 T1 0.043% 0.035% 0.030%Rater 3 TI 0.083% 0.078% 0.070%Rater 1 T2 0.047% 0.032% 0.028%Rater 2 T2 0.050% 0.026% 0.025%Rater 3 T2 0.052% 0.023% 0.023%

To determine which of the improvements in the individual rater's estimates betweenthe first and last estimates was statistically significant, two sided t tests were performed.A probability of P < .05 indicates that a less than 5 in 100 probability exists that thechange in estimate occurred by chance. A probability of P< .01 indicates that a lessthan 1 in 100 probability exists that the change in estimate occurred by chance. Aprobability of P <.001 indicates that a less than 1 in 1000 probability exists that the

33

change in estimate occurred by chance. If the P value is very small it is to be concludedthat the BAC's are more correctly estimated after the test with the smaller absolutedeviation. Large P values indicate that the size of the absolute deviation is due tochance.

, ean t valuieRater 1 -1.49Rater 2 -2.341

Rater 3 -1.44

Team '2Rater 1 -3.543

Rater,2 -4.333

Rater 3 -4.693

Although all of the raters improved the accuracy of their estimates the improvementsin raters I and 3 on team 1 were not statistically significant.

In order to determine if the improvement in individual rater's estimates between theobservation based estimate and the on boat nystagmus based estimates were at-tributable chance t tests were performed. The differences between the first estimateof BAC and the fourth estimate of BAC indicated:

Teaml valueRater 1 -2.781

Rater 2 -1.21Rater 3 -0.20Tcan2Rater 1 -3.252

Rater 2 -4.353

Rater 3 -4.933

Although all of the raters improved the accuracy of their estimates, the improvementsin raters 2 and 3 on team 1 were not statistically significant.

To determine if the improvement in each of the individual rater's estimates betweenon boat testing and the cumulative estimate was statistically significant, t tests were

1 P < 0.05

2p < 0.013p < 0.001

34

performed. The differences between the fourth estimate of BAC and the sixth estimateof BAC indicated:

Rater 1 1.68Rater 2 -2.66Rater 3 -2.03

Rater 1 -1.76Rater 2 -0.44Rater 3 0.00

Only two of the six raters demonstrated a significant improvement (p<0.05%1 ) inestimating BAC from the fourth (on-boat) to the sixth (on-land) estimate (Team 1,Rater 2 and 3). It should be noted that improvements in1 accuracy observed betweentest one and test four by these two raters were not statistically significant.

In summary, the results indicate that the accuracy of the officers to estimate BAC wassignificantly improved by use of the FST battery. The greatest amount of improvementin estimating BAC occurred at the fourth estimate, i.e., following the on boat testing.Small but statistically significant improvements in estimating BAC were evident for twoof the six raters between the on-boat and on-land testing, suggesting that the additionof on-land testing may be beneficial to some officers.

Regarding the use of the HGN both on the boat and on land, it was found that on theboat the HGN contributed to a significant improvement in the estimates of the raters.However, administering the HGN a second time, i.e., on lard, did not result in asignificant improvement in the overall accuracy of the raters. In some states the HGNis administered twice, once on the boat and once on land, in order to give the suspectthe benefit of the doubt. For example, in the court situation the defense might claimthat the results of the HGN on the boat are not admissible as evidence, since theofficer's scoring or the suspect's performance may have been adversely affected by themotion of the boat. The HGN was, therefore, given twice in this study in order todetermine if there was a significant improvement in the raters' estimates attributableto administering the HGN a second time, but this time on land. No significant changein the officers' estimates occurred. Therefore, the results of this study indicate that itis not necessary to administer the test again on land in order to improve the accuracy

I P <0.052 P <0.01

3 P <0.001

35

of estimite proidcd by the I IGN. However, because the on-land HGN test did notresCult in a decrease in acclracy an officer night want to give the I IGN again on land(followrin- on boat I IGN) in order to support (substantiate in Court) the tIGN resultsobtai ned on the boat.

4.3 Arrest/Release Decision

While the correlation between the estimates achieved using the FSTs and the PBTindicate overall accuracy, a more practical question is how well did the use of the FSTprocedures aid the officers in making correct decisions to arrest or release subjects atspecified criteria levels of arrest and conviction?

In order to explore use of the field sobriety tests in making arrest,/release decisionsindices of discrimination were prepared. These indices dezcribe how well an officercan estimate whether an individual's BAC is greater than or less than a set criterion.Two main indices were used: the "True Arrest Index" (TAI) and the "True Release Index"(TRI). The compliment of these indices are the "False Arrest Index" (FAI) and the"False Release Index" (FRI).

These terms are defined as:

True Arrest Index - the proportion of individuals with BACs greater than orequal to a set criteria, who were judged by the rater as having BACs greater thanor equal to that criteria.

False Release Index - the proportion of individuals with BACs greater than acertain criteria, who were judged by the rater as having BACs less than thatcriteria.

False Arrest !ndex - the proportion of individuals with BACs less than a setcriteria, who were judged by the rater as having BACs greater than or equal tothat criteria.

True Release Index - the proportion of individuals with BACs less than a certaincriteria, who were judged by the rater as having BACs less than that criteria.

The true arrest index and the false release index sum to 100% of all subjects with BACsat or above the set convict point criteria and similarly the true release index and falsearrest index sum to 100% of all subjects below the convict point. In calculating theseindices, the BAC estimates of the three raters estimating each individual subject's BACwere summed and divided by three in order to obtain a mean estimate for each subject

36

rather than determining the true release and true arrest index for each rater andaveraging the indices over each team. This technique was used because the BACestimates meet the requirements for parametric analysis while the arrest data do not.

In order to determine if the officers' true arrest and true release indices improved withdifferent criteria the following BACs were used to compute the indices: 0.08%, 0.10%,and 0.12%.

Indices were computed to simulate the following conditions (these indices are foundin tables 4a,b,c,d,e):

1. identification of all individuals with BACs equal to or greater than 0.10%C whenthe conviction criteria is equal to or greater than 0.10%. (table 4a)

2. identification of all individuals with BACs equal to or greater than 0.08%when the conviction criteria is equal to or greater than 0.08%. (table 4b)

3. conservative identification of all individuals with BACs equal to or greaterthan 0.10% by arresting only individuals who were estimated to have BACsequal to or greater than 0.12%. (table 4c)

4. conservative identification of all individuals with BACs equal to or greaterthan 0.08% by arresting only individuals who were estimated to have BACsequal to or greater than 0.10%. (table 4e)

5. liberal identification of all individuals with BACs equal to or greater than0.10% by arresting individuals who were estimated to have BACs equal to orgreater than 0.08%. (table 4e)

4.3.1 Criteria for Arrest at > = 0.10% (equal to or greater than 0.10%) Criteria forConviction at > = 0.10%

Figure 9 depicts the decision indices (percentage of correct decisions based on teamaverages) for the scenario in which the conviction criteria was equal to or greater than0.10% and the officers' criteria for arrest was equal to or greater than 0.10% (condition1).

Table 4a lists the overall arrest/release decisions aggregated over all teams for each ofthe six estimates of BAC under condition 1. By the last BAC estimate (followingadministering the entire test battery) the officers would have arrested 75% of thesubjects having

37

C) C 0 0~\O 0\ N-co C) C ~0\0O \o Cl -() C: 0\O\0 r-c

> "C 0 -i C) > 4 CO1 0 ~ r >~ CO 0 C)

--q:: Cl) LnL H/\~ :j 0 r-4

m)C Lr C)) En CU I r-oococo I (n I

E t ) rz C ) ri C)4

O1 75 0000 c \ o r l 000 0 0

::Ic ,- A N COl r_ ' ro\ \ r- N (n C) 0) N C

41J co 41 1-:cocc) 4~J C N mU)O (Dc 0 (n 0

4-j 4J"1 0 U)\0- 0\0 N5c ) 40OW OU Jj\oO

Z( LlO 0 N CDO :J (o CZ Cl 0)r-1-4 M (-4O0 0 (~Nr:40 co C' 4- : 0 O "T.) t . 0 0 01

(A0 IIC U ON m~ f U) 0A 41 C Un m A 41 Co co A 4-J C -1 O

Cu w) 0u U) 0 CU U)c 0 r-0 < >10< >40 >

Ua U) U- 4J .'-4 4J) -41

.00 000 00 -c a J000\ 4 w 4 \ \ -.> LoC a O T~~ 0> U) CCN r- C 0> M) cO o C

r. 0 : - 0 r- 0 Zr~ rC 0 0 OOa) 0 0 *. 0 z 0. a)0 Z 0.mu) U NO m) u U) IA 0

0)- C) -N 0)- 04 04 rC) O 0a) 4o\ w W

Ci -'-4 CP 0 \

I .4-i 0\0O N\0 co 4-) o' NN0 0\0 4- .JOe 0\ -co-W A m r o NO C4(\ .A fu (C'4 m O f 4J A Q) 0 a%~ N''Mu) 4J a L)O (n u 4J 9 nw nu 4.) ~00a)< .1 C 0< .1-4 Q) l< -4

C)' 0 ~ w.j Q) r- w) 9:

41) 41 41)En U) *-U) CA4U

'4) o~' .a a8~) m~'

a) 4 a)0 4 < -> >

0\ 0\ 0-c \ \ : \ \ -0o NO 0o 0.r m m n 000 0) N3-I LO -- He m -- 400

41 Q10 4.1 - 4. 0 4*- 40 H- (a co- 00 0

>0 ON 0

41 ul 4.1 En .4-) IA

'0 H0 H0

0.. 0 U)l 0. 04 4) 0. 0) wo

00 0 0 0 0 0C 4 -l -r- C$.,4 1-OC ~4 Tl

0 w Q .0 11 0 w ) .11 0 $ 0) .0113 A ~ -- 4 <1 :JA *-< ~A

10~ ~ Q a) U) 0 ) U)0U)-4 I-4 0(a U) '-4 *- 4) -4 *OO H

0 -E ~ < ) E-4 0E--4 E- ) E-4 E-E-4 H U

38

V.0 > C) C 7 0,0 0\0 co

0 >00CC> "0 or ° CD° . " oOoo

L- I >

S ,o o o\0 - rl (n 75 0\0 o\0 r 0

SAC I) (0 -1C

4-) 4-) COr

no 00\ (0 0\rHC

co 4J o\O 0\0 k0 a 0 r 0 M 4- 0\0 0\o

0-, o3 (a r - C) .H --4 :: (a -- Q0 0) (N

II a M' , :1 I 0 cO

a (0 ) (0U) (n

* 4-) 4O .)

•a.. ) 41 0\0 0\0 r,- r- 4J.- IJ o\o do~o r-o0 :3 "; 0) c m I 0 o I r-

a) Z0 a0 *0 •II \O \O t II •, Oco C

En u La- U Nr

a)) 1(2a* \H ONi a)-4 0\0 0

Q)0 ca 00

0> 0\0 H\ 4- 0\ 0\ ,0 oc0

O0 ' 0Q H

0)0CO . C4 co0o

4 o- o l- -

5i 0, EnI

00 0)c

~z4 H r,03 - -*

4 O\ rr A .(4 Od) V C A39 (0 0)c

C2 5l0) )O

411 U) 4-4U)

HO) 4J H)0\C05- W m00 4

w-5 r5-450E4 n uH 0)) QH0

0 0Q1 0 aoH 00 a) cl 0 AC(

E44 &-o r-4 ~-

Im~f 4-lU)U)E-4E4H O- - - - - - - - - - - - - -

039

A-4q

-F-

-..... .....- lo-1 0

40>

BA's of equal to or greater than 0.10% and would have falsely released 25% of thesubjects having BACs of equal to or greater than 0.10%, i.e., estimating that their BACswere less than 0. 10%'c. The officers would have correctly released 86% of the subjectshaving BACs less than 0. 1() and would have falsely arrested 14% of the subjects havingBACs less than 0. 10%c.

In the two previously mentioned studies of field sobriety testing in the highway situation(Burns and Moskowitz and Tharp, Burns, and Moskowitz), the total number of correctarrest/release decisions made by the officers was 76% and 81%, respectively. Forpurposes of comparison, the percentage for the total number of correct decisions wascomputed for the current study. Of 97 subjects, the officers correctly classified 21 truearrests (persons with BAC > =0.10%) and 59 true releases (persons with BAC<. 10%). This represents 80 correct classifications out of 97, i.e., 82% correct decisions.

Table 5 lists the data used in this calculation.

Table 5. Decision Index for All Raters' Estimates of BAC

Arrests Releases All

True True TrueNumber of Decisions Arrests Releases Decisions

21 59 80

False False FalseNumber of Decisions Arrests Releases Decisions

10 7 17

All All AllNumber of Decisions Arrests Releases Decisions

31 66 97

Percent of All Correct Decisions = 82%

These results were achieved through use of the entire field sobriety test battery, i.e., allon-boat and on-land tests. This battery includes the three tests which were found to be

41

the -best abbreviated battery" of field sobriety tests which were studied for use in thehiglhway situation and additional FSTs used in the marine environment.

In the 1977 Burns and Moskowitz (B.&M.) study of "Psychophysical Tests for DWIArrests", they found that at the 0.10%c criteria the officers correctly decided to arrest84% of the cases having BACs equal to or greater than 0.10%, and for less than 0.10%/'-

they made the correct decision to release 73% of the time. By computing the mean ofall correct decisions made by the officers, it was found that the officers made correctdecisions in 76% of all decisions made, i.e., correctly classifying 76% of the subjectswith regard to a BAC of 0.10%.

In the 1981 Tharp, Burns, and Moskowitz (T,B.&M) study of "Development and FieldTest of Psychophysical Tests for DWI Arrest", they found that at the 0.10% criteria theofficers correctly decided to arrest 64% of the cases having BACs equal to or greaterthan 0.10%, and for less than 0.10% they made the correct decision to release 88% ofthe time. By computing the mean of all correct decisions made by the officers, it wasfound that the officers made correct decisions in 81% of all decisions made, i.e.,correctly classif'ying 81% of the subjects in regard to BAC of 0.10%. The results of thecurrent (Yorktown) study are compatible with the findings of the two previouslymentioned studies by Burns and Moskowitz, and Tharp, et al. The Table 6 summarizesthe findings of the three studies.

Table 6. Comparison of Marine and Non-marine FST IndicesNys _ Cur. B. & M. T.,B.& M.

True Arrest Index 52% 75% 84% 64%False Release Index 48% 25% 16% 36%True Release Index 94% 86% 73% 88%False Arrest Index 06% 14% 27% 12%

In addition to comparing the three studies with regard to percentage of correctdecisions made by the raters, a comparison was made of the Arrest Error Rate. TheArrest Error Rate was calculated in order to compare this work with the two earlierstudies mentioned. However, the particular values found may not hold in real worldsettings. The Arrest Error Rate is the number of false arrests divided by the numberof total arrests. While the False Arrest Index identifies the percentage of subjects whohad BACs of less than a set criteria but were judged by the rater as having BACs greaterthan or equal to that criteria, the Arrest Error Rate identifies the percentage of subjectserroneously arrested based on all subjects arrested.

42

The Arrest Error Rates of the raters of the three studies are listed in table 7.

Table 7. Arrest Error RatesStd y False Arrests/Total Arrests Arrest Error Rate

Yorktown 10 out of 31 3 2cB. & M. (1977) 47 out of 101 46%cT,B.& M. (1981) 38 Out of 118 3 20 c

The results of the current study are consistent with the findings f the two previouslymentioned studies by Burns and Moskowitz, and Tharp, et al. In the Burns andMoskowitz study the high Arrest Error Rate was attributed to the fact that this was alaboratory studv and was not indicative of the real arrest situation in the field. Theystated that, officers in the field are reluctant to err in the directioii J, falc arrcsts, andobservations indicate that the most common error probably is a "false negative". A"false negative" refers to the situation in which an individual's actual BAC is equal toor above a set criteria and the officer estimates that the BAC is less than the set criteria.This would result in the officer releasing the suspect who was "legally" intoxicated. Theofficers in the Yorktown Study confirmed the prevalence of this type of decision-making. Basically, when in doubt, the officer would rather err in the direction ofreleasing a guilty person rather than arrest an innocent person.

In both the Tharp, et al. study and the Yorktown Study the Arrest Error Rate was 32%.While this is lower than the earlier Burns and Moskowitz findings, an Arrest Error Rateof 32% is quite high. It essentially means that the officers would have arrested oneinnocent person in every three arrests. It is believed that this high Arrest Error Rateis due to the fact that the officers may have decided to arrest a greater number ofindividuals than they would have arrested in the field, since these "arrests" were notactual arrests but were part of an experimental study. Since there was no consequencefor either the officer or the subject, as a result of an c.rroneous decision to arrest, theofficers may have been inclined to decide to arrest rather than to release. In the fieldthis would not be the situation and fewer arrest errors would occur.

It should also be noted that in the present study, the decision to arrest cannot be directlyprojected to actual performance, because none of the subjects reached very high BACs(0.16%) the officers' decision processes were made slightly more difficult. In actualpatrols officers will encounter a small number of extremely intoxicated individuals.These individuals are presumably more readily detected as intoxicated. Therefore,under actual conditions the officer's decisions regarding arrest may be improved.

43

4.3.2 Criteria for Arrest at > = 0.08% Criteria for Conviction at > = 0.08%

Figure 10 depicts the decision indices for the scenario in which the conviction criteriawas equal to or greater than 0.08% and the officers' criteria for arrest was equal to orgreater than 0.08%. Table 4b lists the overall arrest/release decisions for al! officersfor each of the six estimates of BAC. By the last BAC estimate (following administeringthe entire test battery) the officers would have arrested 86c/( of the subjects havingBACs equal to or greater than 0.08%, Inversely, the officers would have falsely released14% of the subjects having BACs equal to or greater than 0.08%. The officers wouldhave correctly released 80% of the subjects having BACs less than 0.08%. Inversely,the officers would have falsely arrested 20% of the subjects having BACs less than0.08%. It follows logically that if a lower arrest/conviction criteria is set, more subjectsare subsequently arrested, however, it is at the cost of a greater number of false arrests.

4.3.3 Criteria for Arrest at > = 0.12% Criteria for Conviction at > = 0.10%

Figure 11 depicts the decision indices for the scenario in which the conviction criteriawas equal to or greater than 0.10% and the officers' criteria for arrest was equal to orgreater than 0.12%. Table 4c lists the overall arrest/release decisions for all officers foreach of the six estimates of BAC. This scenario represents a situation in which theofficer is being conservative in making the decision to arrest. Often officers in the fieldwill make such a decision in order to increase the probability of a true arrest anddecrease the probability of a false arrest. Officers often arrest at an estimated BACwhich exceeds the BAC needed for conviction, since a significant amount time is lostin transporting the suspect to the site of the breathalizer unit. During this transporttime the suspect's BAC level can decrease, for example, approximately 0.015% perhour elimination rate dependent on rate of absorption and body weight. Therefore,the officer feels more confident when arresting the suspect who is estimated to be0.02% above the conviction criteria.

In this scenario, by the last BAC estimate (following administration of the entire testbattery) the officers would have arrested 39% of the subjects having BACs equal to orgreater than 0.10%. Inversely, the officers would have falsely released 61% of thesubjects having BACs equal to or greater than 0.10%, i.e., estimating that their BACswere less than 0.10%. The officers would have correctly released 97% of the subjectshaving BACs less than 0.10%. Inversely, the officers would have falsely arrested 3%of the subjects having BACs less than 0.10%. Logically, since the officers are beingmore conservative in their decisions to arrest, many fewer subjects would have beenarrested than in the two previous scenarios, i.e., 39%. When being conservative theofficers indeed make fewer arrests, however, the benefit appears in

44

F-

-NO

I ~ K I

- Q - A A

1-45

0-

z

-4

-C,

-

H H

K K K4 ~ x 7/ 'K -

,c,\' A~% K/\A / -K 'KKY"/~ 'K~ 'K<'K</7'Kkx/ -

41 \-<N N <>7>>~>~

-- xttv x,<xx$~y y <A

L A -

«Kkj:L AK§K& <K 'K I

II \. >~. f.ZI£

Co

~

.- Q

0

t

Cr Co

0

Co

o 0 0 0 0 0o CO CQ

SIIOISIDDQ 4W1L1103 %46

a true release rate ot,)71(. Therefore, more of the guilty are being released but less ofthe innocent are being arrested. In addition, the arrest rate appears to be dispropor-tionate low, since in this study very few subjects actually had BACs equal to or above0.121 Once the officers suspected this fact, they might have concentrated theirest inates closer to a maximum of 0.10V%( BAC

4.3.4 Criteria for Arrest at > =0.10% Criteria for Conviction at > =0.08%

Figure 12 depicts the decision indices for the scenario in which the conviction criteriawas equal to or greater than 0.08%'2 and the officers' criteria for arrest was equal to orgreater than (). 10%. 1thle 4d lists the overall arrest/release decisions based on the sixFST estimates of BAC. By the last BAC estimate (following administering the entiretests battery) the officers would have arrested 68% of the subjects having BACs equalto or greater than 0.08%1 .Inversely, the officers would have falsely released 32% of thesubjects having BACs equal to or greater than 0.08%, i.e., estimating that their BACswere less than 0.08%. The officers would have correctly released 90% of the subjectshaving BACs less than 0.08%. Inversely, the officers would have falsely arrested 10%of the subjects having BACs less than 0.08%. As in the previous scenario, at this setcriteria the officers are taking a conservative position. When doing so, less true arrestsare made and more true releases result.

4.3.5 Criteria for Arrest at > = 0.08% Criteria for Conviction at > = 0.10%

Figure 13 depicts the decision indices for the scenario in which the conviction criteriawas equal to or greater than 0.10% and the officers' criteria for arrest was equal to orgreater than 0.08c. This is a situation in which the officers, in order to increase thenumber of true arrests, have lowered the BAC arrest criteria. Therefore, they wouldbe arresting more suspects and releasing less suspects. However, they would bearresting some suspects who were actually below the conviction criteria, i.e., arrestinginnocent persons. This is not a strategy which an officer usually would adhere to, sinceit is the officer's goal to arrest the guilty and release the innocent. However, an officermight wish to begin the OUI investigation with the strategy of detaining a suspect forfurther testing and investigation rather than for arrest, when the officer has estimatedthe suspect's BAC to be slightly below the conviction criteria. Once further testing hadbeen conducted, the officer would then make the arrest/release decision based on theadditional information and would use a criteria for arrest which is consistent with thecriteria for conviction.

47

0xL6)

-48

444

X 0X

0 00

-4 49

Essentially, the officer is using a finer filtering or screening process as the investigationproceeds. Tie officer begins by detaining all Suspects who are estimated to be at orabove the conviction criteria 08+.. Table 4e lists the overall arrest/release decisions for

all officers for each of the six estimates of BAC. By the last BAC estimate (followinuadministering the entire test battery) the officers would have arrested 89% of thesubjects having BACs equal to or greater than 0.10%. Inversely, the officers wouldhave falsely released 1 1c of the subjects having BACs equal to or greater than 0. 10(',i.e., estimating that their BACs were less than 0.10%'. The officers would have correctlyreleased 72% of the subjects having BACs less than 0.10c. Inversely, the officerswould have falsely arrested 28% of the subjects having BACs less than 0. 10cc. There-fore, while more guilty suspects were arrested, more innocent suspects were alsoarrested.

4.4 Considering the Two Stage Test Process

The above examination of arrest/release decisions applies to the situation in which theofficer has given the entire test battery. This would be analogous to the officer in thefield asking all suspects (regardless of the officer's estimate of BAC) to return to shorein order to perform the on-land testing. However, in actual OUI detection, the officerwould probably give the on boat tests to a suspect and reserve on land testing only forsuspects who the officer estimated to have a BAC equal to or greater than the legalconviction criteria.

In the highway situation it is routine for the officer to administer on land balance tests.The officer simply asks the suspect to step out of the car but the highway officer doesnot have to transport the subject in order to perform the testing. The marine lawenforcement officer must take the suspect to shore for further testing, a process whichis quite time consuming. The officer needs to be fairly confident in his suspicions thatthe person is intoxicated before taking him/her to shore. It is, therefore, necessary thatarrest/decision accuracy be examined for on-boat testing alone. The question is: Howeffective is the on-boat testing in aiding the officer to identify operators to be arrestedfor OUI?

As previously mentioned, Figure 9 and Table 4a refer to the decision indices for thescenario in which the conviction criteria was equal to or greater than 0.10% and theofficers' criteria for arrest was equal to or greater than 0.10%. By the fourth BACestimate (following the last on boat test, i.e., horizontal gaze nystagmus) the officerswould have arrested 52% of the subjects having BACs of equal to or greater than 0.10%.Inversely, the officers would have falsely released 48% of the subjects having BACs ofequal to or greater than 0.10%, i.e., estimating that their BACs were less than 0.10%.The officers would have correctly released 94% of the subjects having BACs less than

50

0.10%. Inversely, the officers would have falsely arrested 6% of the subjects havingBACs less than 0.10%.

When these results are compared to the arrest/release decision indices which resultedfrom the officers administering the entire test battery, based on the results of on boattesting alone officers arrested a fewer number of suspects with BACs equal to or greaterthan 0.10%, i.e., 52% versus 75%.

Conversely using on boats testing officers falsely arrested a fewer number of suspectswith BACs less than 0.10%, i.e., 6% versus 14%.

If on boat testing is used with a conservative arrest criteria, the results still appear tobe worthwhile. This scenario would involve the officer being reluctant to bring a boaterto shore for arrest unless he or she had great certainty that the suspect was legallyintoxicated (the officer estimates the suspect's BAC is higher than the convictioncriteria). For example, if the conviction criteria is 0.10% the officer might only bringsuspects to shore for arrest with estimated BACs of 0.12% or greater.

Using the data from the current study, this would mean that after on boat testing theofficer would have arrested 11% of the suspects having BACs greater than or equal to0.10%. This would result in the officer detaining approximately one intoxicated suspectout of nine suspects having BACs greater than or equal to 0.10%.

Arresting only one guilty suspect out of nine may appear very inefficient. However, itshould be considered that the limiting factor on how many suspects can be arrested andconvicted is to a large extent the time required in the administrative proceduresrequired to obtain a conviction. An OUI arrest often is a lengtl'v and costly process.On boat testing takes approximately fifteen minutes. If an arrest is to be made, theofficer must make arrangements for securing the suspect's boat. Even in states whichuse breath or other chemical tests equipment may be located at distances which mayrequire anywhere from fifteen minutes to an hour transport time. During all this timethat officer is off the water and, therefore, not available for safety patrol. Finally inmany jurisdictions the officer must devote considerable time to court procedures whichlimit the time he or she is available for patrol.

Using a conservative criterion the officer would be almost certain of a conviction.Considering the time required to make an arrest and the inconvenience to the boaterwho is falsely arrested, the use of the on boat tests with a conservative set point wouldincrease the officer's confidence and make him or her vigorously proceed with thearrest procedure.

51

Analysis was performed in order to investigate possible strategies which can be used intestino in two stages. The arrest/release indices were calculated for various set criteriafor on boat and on land testing. For example, instead of viewing the on boat testingcriteria as arrest criteria, it can be considered as criteria for further investigation.Specifically, consider the scenario in which the officer decides that if as a result of onboat testing, he or she feels that the suspect's BAC is approximately 0.06% that suspectwill be taken to shore for on land testing. Once on land the criteria for arrest will be0.10%, the legal conviction level. In lowering the BAC criteria on the boat, the officerwill be bringing more suspects to shore for on land testing than he or she would if onlysuspects estimated to be at 0.10% BAC were returned to shore for on land testing. Thiswould be a strategy which result in a greater number of true arrests, but would alsoresult in a lesser number of true releases. It would also result in a greater inconveniencefor both the boater and the officer, i.e., time spent in testing. Table 8 lists a few suchscenarios in which different criteria are set on the boat and on land.

Table 8The Two Stage Decision Process

BAC CriteriaOn Boat On Land TrueArrt True Releases0.10% 0.10% 48% 94%0.08% 0.10% 59% 87%0.06% 0.10% 70% 86%

The above table represents the decision process regarding 96 subjects. (One of the 97subjects was omitted from the analysis, since on boat HGN was not performed for thatsubject due to technical difficulties with the boat on which he was a passenger.) Thisinformation is presented so that law enforcement officials may use these results as anaid in planning their arrest strategy. If their goal is to obtain the greatest number ofarrests then lowering the criteria for further investigation is a reasonable approach.However, if their goal is to minimize the number of false arrests (maximize truereleases), then maintaining an equal criteria for further investigation and criteria forarrest on land is a reasonable approach.

52

5 CONCLUSIONS

The objectives of this study were to determine if field sobriety tests were invalidatedby the exposure of the individuals tested to the marine environment and to developestimates of the effectiveness of a field sobriety test battery in aiding the marine lawenforcement officer in identifying intoxicated boat operators. The data indicate that:

A. The accuracy of the field sobriety tests, when used on subjects exposed torecreational boating conditions, was as good as the accuracy of such tests whenevaluated under simulated highway conditions.

B. The administration of the entire test battery improved the accuracy of theofficers' estimates of BAC over estimates based on observation of and conver-sation with the suspect. As noted below, even the on-board Horizontal GazeNystagmus test resulted in a substantial improvement in accuracy over obser-vation and conversation.

C. The Horizontal Gaze Nystagmus procedure appeared to provide the raterswith the most accurate information of any single test. The major improvementin accuracy (of the combined estimates of all raters) occurred after the fourthBAC estimate which was based on-boat testing using the Horizontal GazeNystagmus test).

D. No significant difference in the accuracy of the combined estimates of allraters was found as a result of on land testing subsequent to the on boat testing(though two of six individual raters demonstrated small but statistically sig-nificant improvements in the accuracy of their BAC estimates as a result of theaddition of on land testing).

53

6 RECOMMENDATIONS

The following recommendations cover the use of Training, Chemical Testing, On-boatField Sobriety Testing (FST), On-Land (FST), and the Two Stage testing process:

Training:* It is recommended thaL inarine law enforcement officers be thoroughly

trained in the use of field sobriety tests and in the use of portable breathtesting equipment , especially in the use of these tests on the water.

- Officers using FSTs should undergo a training course and in stateswhich have a certification process be certified for use of the HGN.Once the officers have been trained they must have the oppor-tunity to practice use of the test in the field to keep this skillcurrent.

- Formal training in the use of portable breath testing equipmentis also necessary. Such training is required not only to ensureaccurate testing but also to ensure that the officer adheres toprocedures that result in obtaining admissible evidence.

Chemical Testinge In states where permitted, portable breath testing units should be used to

make an preliminary arrest/release decision subject to confirmation by otherchemical testing. PBTs provide a quick and reliable reading of BAC.

- If chemical testing is to be used to establish evidence of legalintoxication, time is critical. The suspect's BAC level decreasesas alcohol is eliminated from the suspect's system. It is imperativethat once the officer is confident that the person is intoxicated(based on the on boat testing), that the officer makes every effortto have the suspect agree to a chemical test of intoxication andensure that the test is performed as soon as possible.

Field Sobriety Testing9 In those jurisdictions where chemical or breath testing to determine BAC is

not possible, on-boat field sobriety testing provides critical aid to the officerin making arrest-release decisions. Even in cases where chemical tests areused to support boating intoxication convictions, field sobriety testing cansubstantially help the officer in making arrest/release decisions.

54

On-Boat FST" The on-boat FST procedure should include the use of HGN. As noted above

this provides the officer with the most accurate information of any on-boatprocedure

" It is recommended that marine law enforcement officers using FST proce-dures, where possible, use the methods of interview, and observation, as wellas tHGN testing. Each method can provide additional information on thesuspect's level of intoxication.

" The observation portion of the FST should include the use of a checklistsimilar to that used in this study to refresh the officer's memory regardingpoints of observation. Through a structured interrogation the officer gainsimportant information on any impairment of the suspect's speech and logic.

" If the officer suspects that the person is legally intoxicated the officer canproceed to performance tests, such as the alphabet recital, hand pat, finger-to-nose, and finger count as used in this study. Other similar performancetests may also be of value. The most essential element in performance testingis that they provide a structured situation in which the officer can observe adefined behavior.

" It is recommended that where on-boat testing is possible the HGN be givensubsequent to the performance tests. Although as noted above the HGNappeared to be the most useful procedure, because this study assessed theeffectiveness of an entire test battery it can not be proven that the administra-tion of the HGN alone would result in the same accuracy as the administrationof all the on boat tests.

As a practical matter it would be highly unusual for an officer toapproach a suspect and immediately administer the HGN. Inter-view and observation is always the first step in a criminalinvestigation. The use of performance testing provides time forthe officer to further observe the suspect's abilities and possibleimpairment.

55

On-Land FST* Unless "Two-Stage Testing (see below) is to be employed it does not appear

necessary for the officer to administer the HGN a second time, i.e., on-land.This study did not find significant differences between the scores obtainedwith the on-boat HGN and the on land HGN.

- On land testing is time consuming (ten to fifteen minutes) andcan be left to the discretion of the officer, if permissible underdepartmental procedures. The officer may wish to give the HGNon land in order to confirm the score obtained on the boat.

- Although the use of the One-Leg Stand and Walk and Turn didnot result in a significant improvement in the accuracy of thecombined BAC estimate of all officers, two of the six officers didshow small but significant improvements in accuracy. These testsmay provide additional information to some officers.

Two Stage testing* In addition to using on land testing in order to confirm the results of on boat

testing and to further substantiate the arrest-release decision, on land testingcan be used be used to supplement on-boat testing through the two stagedecision strategy mentioned in the section 4.0.

- By using different BAC criteria to identify suspects for furtherinvestigation (during on-boat testing) and criteria for arrest (dur-ing on-land testing), the officer can increase or decrease the truearrest rate. But it is important to remember that as the true arrestrate increases, the false arrest rate also increases.

- Where possible a two stage procedure is recommended, forexample: If after testing on the boat the officer is reasonablyconvinced that the suspect's BAC level exceeded the legal BACconviction criteria of the state (in a state which has a per se law)and/or the tests indicate impairment which may pose a danger (instates with no per se law) the officer would bring the subject toshore and use the on-land testing for confirmation.

56

APPENDIX A

INDEX TO APPENDIX ITEMS

Page

1. Scoring A-2

2. Age Distribution of Participants A-10

3. Participation Announcement in "Plan of the Day" A-lI

4. Telephone Screening Instrument A-12

5. Alcohol Questionnaire A-13

6. Years of Experience in Marine Law Enforcementand OUI Arrest A-20

7. Dosing Tasks A-21

8. Nominal Dosing Levels A-24

9. Actual Dosing Levels A-25

10. Medical Report A-28

11. Subject Briefing A-29

12. Informed Consent Form A-30

A-I

Item I

SCORING SHEET

Subject #: Officer:Date: Time:

INTERVIEW:

Officer introduces self to subject. Asks subject for some identification and askssubject's name, address, and phone number, while subject is looking for ID.Officer asks: Response:Without looking at your watch,what time is it?What is today's date?What day of the week is it?When did you last eat?What did you eat?Where did you eat?When did you last sleep?How long did you sleep?Are you ill or injured?

OBSERVATIONS

CLOTHES

Condition: C DISORDERLY C3 DISARRANGED C3 SOILED C MUSSED 0 ORDERLY(Describe)

BREATH Odor of Alcoholic Beverage: [ STFONG [ MODERATE [ FAINT C3 NONE

ATTITUDE C EXCITED C HILARIOUS C TALKATIVE C] CAREFREE 0 SLEEPY C PROFANITY[] COMBATIVE C INDIFFERENT - INSULTING C COCKY C2 COOPERATIVE 0 POLITE

COLOR OF FACE 0 PALE C FLUSHED C3 NORMAL C OTHER

EYES C BLOODSHOT C WATERY C NORMAL

PUPILS C NOT EQUAL SIZE C CONTRACTED 0 DILATED C NORMAL

UNUSUAL ACTIONS ED HICCOUGHING [ BELCHING C VOMITING C FIGHTING C CRYING :: LAUGHING C NONE

SPEECH C NOT UNDERSTANDABLE " MUMBLED C SLURRED C MUSH MOUTHED C: CONFUSEDC THICK TONGUED C STUTTERED C ACCENT C FAIR C: GOOD

INDICATE OTHER UNUSUAL ACTIONS OR STATEMENTS. INCLUDING WHEN FIRST OBSERVED:

BAC ESTIMATE:

A-2

Subject ;:

1. ALPHABET RECITAL

INSTRUCTIONS:

Please recite the alphabet A through Z.

SCORING:

SatisfactoryRecited slowly_Thick and slurred speech

Omitted lettersUnable to perform

2. HAND PAT

INSTRUCTIONS:

Watch what I do so you will be able to do the same thing. Don't begin until I

tell you. (officer demonstrates.) Hold both palms out, facing up. Keep the left

hand stationary and clap your palms together and count ONE. Turn your righthand over and clap the back side of your palm and count TWO. Keep counting(still demonstrating) ONE, TWO until I tell you to stop. As you count I want youto increase your speed.

Do you understand? Ready? Begin.

Have the subject count ONE, TWO about 10 times.

SCORING:

SatisfactoryFailed to alternatepalm and back of handHit side of handSlowUnable to perform

BAC ESTIMATE:

A-3

Subject =:

3. FINGER TO NOSE

INSTRUCTIONS:

Subject is seated. Hands at his sides.

W atch what I do so you will be able to do the same thing. Don't begin until Itell ."ou. (Officer demonstrates.) I want \ou to close your eyes. put .%our armsdown to your sides. tilt your head back slightly and when I say RIGHT, bring thetip of your right index finger to the tip of your nose. Then return \. ur armdown to your side. W hen I say LEFT, bring the tip of \our left index finger tothe tip of your nose. Then return 'our arm to the side.

Do you vinderstand? Read\?

Interrupt if there is significant deviation from the instructions. Repeatdemonstration. Give second trial or discontini'e.

Have subject perform RIGHT and LEFT for 2 or more trials.

SCORING:

Sure, accurateSlow but accurateUncertain, fumbled but touchedRequired repeated instructions/demostrationsDid not return arm to startingpositionTouched with second and thirdfinger joint, or with otherfinger, or with entire handMissed completely

A-4

Subject =:

4. FINGER COUNT

INSTRUCTIONS:

Watch what I do so you will be able to do the same thing. Don't begin until Itell you. I am going to touch my thumb and finger and count like this.(Demonstrate slovly and with slight exaggeration.) 1-2-3-4-4-3-2-1. Each timeyou do this, do it a little faster than the time before.

Do you understand? Ready? Begin.

Have subject count 1-2-3-4-4-3-2-1 three times.

Interrupt if there is significant deviation from the instructions. Repeatdemonstration. Give second trial or discontinue.

SCORING:

Sure, accurateConfused, started overCounting errorDid not correctly touchthumb to fingerRequired repeated instructionsUnable to perform

BAC ESTIMATE:

A-5

Subject #:

S. HORIZONTAL GAZE NYSTAGIUS

INSTRUCTIONS:

I am going to check your eyes. (If subject wears glasses ask him to remo~e

them.) Please look at the tip of the pencil. Now, keep your head still and follow

the tip of the pencil with your eyes. Only move your eyes, not your head. )o

'ou understand" Watch the tic of the pencil.

(If the subject keeps moving his head, have him cup his chin with his hands.)

SCORING:

Right Eye Left Eye

I. Onset occurs before 45degrees

2. Moderate/distinctnystagmus at extremes

3. Cannot smoothly follow

a moving object

Total Score:

BAC ESTIMATE:

A-6

ON-L\ND ESTING

HORIZONTAL G.-\E N'rtS\G MlS

INSIRI, FIONS:

I am going to check 'our e,,es. (if subject wears glasses ask him to remu'ethem.) Please look at the tip of the pencil. Now, keep ,,our head ,till indfollow the tip of the pencil with your eyes. Only move ,our eyes, not %,',urhead Do ,ou understand? Watch the tip of the pencil.

(if the subject keeps moving his head, have him cup his chin with his hands.)

SCORING:

Right Eye Left Eye

1. Onset occurs before 45degrees

2. Moderate/distinctnystagmus at extremes

3. Cannot smoothly followa moving object

Total Score:

BAC ESTIMATE:

A-7

Subject #:

7. ONE-LEG STAND

INSTRUCTIONS:

I am going to test your balance. Please stand with your heels together andyour arms down at your sides, like this. (Demonstrate position.)

When I tell you to begin, I want you to raise one leg so that .our heel isabout six inches off the ground, and hold that position. Watch the toe of ,ourraised foot and at the same time count from 1001 to 1030...like this.(Assumethe position and demonstrate how you want the subject to stand and count.)

Do you understand? (Do not continue until the subject indicates that heunderstands.) Begin by raising either your left or sour right foot andcounting.

SCORING:

1. Sways while balancing on one leg.2. Moves arms more than six inches to

maintain balance.3. Hops on one leg to maintain balance.4. Puts foot down one or two times during

thirty-second count.(Count this itemonly once.)

5. Cannot do the test. (Puts foot downthree or more times, or loses balance.Score this item five points.

Total Score

A-8

Subject 4:

8. WALK AND TURN

INSTRUCTIONS:

I am going to give you a test to check your balance. Please put your left Cooton the line and then your right foot in front of it with your right heeltouching your left toe. (Demonstrate the position.)

When I tell you to begin, take NINE heel-to-toe steps down the line, turnaround, and take NINE heel-to-toe steps back. In turning around, make %ourturn by pivoting on one foot. Keep it on the line and use your other foot toturn yourself around with several small steps, like this. (Demonstrate) Keepyour hands at your sides at all times, watch your feet at all times, and countyour steps out loud. Do you understand? (Demonstrate again if subject doesnot understand.) Begin, and count your heel-to-toe steps out loud.

SCORING:

I. Loses balance during the instructions.2. Starts before the instructions are finished.3. Stops or pauses for several seconds

while walking.4. Doesn't touch heel-to-toe.

(Leaves more than 1/2 inch.)5. Steps off the line one or two times.

(Count this item only once.)6. Raises one or both arms more than six

inches to maintain balance.7. Doesn't turn correctly or loses

balance during turn.8. Takes more or less than nine steps

in each direction.

9. Cannot do test. (Steps off the line threeor more times, is in danger of falling, orotherwise demonstrates inability to completethe test. Score this item nine points.)

Total Score:

BAC ESTIMATE:

A-9

Item 2

AGE DISTRIBUTION OF BOATING SAFETY PARTICIPANTS

Age No

21 2022 723 624 725 926 727 328 229 830 631 83233 5343536 137 238 339 2404142 1

Total 97

A-10

Item 3

United States Coast Guard Reserve Training Center)

PLAN OF THE DAYPride - Professionalism - Excellence

Capidlait WM. A. DoigUSCGCommanding Officer

BOATING AND ALCOHOL STUDY: Volunteers are needed to participate in astrictLy controlled study concerning the effects of alcohol onboaters. The study will be conducted by a private research grouputilizing RTC fdcilities from 11-22 May 1987. Criteria forvolunrteers: Male, 21-50 years of age, be in good health. Volunteerswill be asked to: Drink measured amounts of alcohol, take a one-hourboar ride, take breath tests, participate in several field sobrietytests. The study will be conducted between the hours of 0700 and[200. (Times for volunteers to report in). Each volunteer will berequirud ro participate for one 8-hour period during the 10-day study.Alt sessions will be medically supervised. After testing,prricipcints will be given food arnd rest prior to being driven home.If inturested, contdct MLF. SchooL, X2194.

A-I1

Item 4

TELEPHONE SCREENINC INSTRUMENT

DATE:

NAME: MALE

ADDRESS: WEIGHT:

HEIGHT:

TELEPHONE NUMBER: DAY: NIGHT:

1. AGE:

If less than 21 or greater than 50, tell them they do not qualify.

2. Are you presently taking any prescription or over-the-counter medication

or drugs?

If yes, what?

3. Do you go out on a boat?

Often Rarely Never

4. Are you prone to seasickness?

Yes No

5. Do you ever drink alcoholic beverages?

Yes No

6. Have you ever been advised by a physician to abstain from or reduce theamount of drinking?

Yes No

If Yes, Why?

A-12

Telephone Screening - Page Two

7. How often do you consume alcohol?

dailyseveral/weekonce/weekseveral/ monthonce/month or lessnever

8. When you drink, do you usually drink

1-2 drinks3-4 drinks5-6 drinksOver 6

Explain to the caller that an initial screening will be necessary. It will involvefilling out a questionnaire that will take 15-20 minutes. This must take placebefore the actual experiment.

Availability to Participate in Study

THANKS A LOT FOR YOUR INTEREST, AND WE'LL GET BACK TO YOU SOON.

A-13

Item 5

Name:

Telephone No.: Day

Evening-

ALCOHOL QUESTIONNAIRE

FIRST PART

1. On a typical occasion when you are drinking distilled spirits (that is,whiskey, gin, vodka or beverages of that sort), how much of that type ofbeverage to you usually consume?(Interviewer: Check one response; if a range of quantity is cited, checkthe lower figure.)

doesn't drink distilled spirits 0one shot 1two-three shots 2four-five shots 4six-seven shots 6eight-ten shots aone pint 10one pint to one fifth 15more than one fifth 20

2. On a typical occasion when you are drinking beer, ale, or malt liquor, howmuch of that type of beverage do you usually consume?(Interviewer: Check one response; if a range of quantity is cited, checkthe lower figure.)

doesn't drink beer 0one 12-ounce bottle 1two-three bottles 2four-five bottles 4one-two six-packs 10more than two six-packs 15

3. On a typical occasion when you are drinking wine, how much wine will youusually consume?(Interviewer: Check one response; if a range of quantity is cited, checkthe lower figure.)

doesn't drink wine 0one glass (3-4 ounces) 1two-three glasses -2four-five glasses 4one bottle 10more than one bottle 15

A-14

Alcohol Questionnaire-Pag-2 2

4. How often do you drink during the morning?

daily 30several/week 25once/week 15several/month 10once/month or less 5never 0

5. How often do you drink at lunchtime?

daily 8several/week 5once/week 3several/month 2once/month or less 1never 0

6. How often do you drink during the afternoon, that is, after lunch is overbut before the cocktail hour begins?

daily 15several/week 10once/week 7several/month 3once/month or less 1never 0

7. How often do you drink at dinnertime, that is. either just before dinner orduring the meal itself?

daily 5several / week 4once/week 3several/month 2once/month or less Inever 0

8. How often to you drink during the evening, that is, after dinner is over?

daily 5several / week 4once/week 3several/month 2once/month or less 1never 0

A-15

Alcohol Questionnaire-Pag,: 3

9. Speaking once again about your own typical drinking occasions, would yousay that you most often drink in a private home, a bar, a restaurant, orsome other place?

private home 1bar/restaurant 2other (specify)

10. Once again in relation to your own typical drinking occasions, would yousay that you most often drink with members of your family, with friends,with barroom clientele, or alone?

family 1friends 2barroom clientele 4alone 8

11. At this point, I want you to think about the drinking you have doneduring the past 12 months. In particular, on how many occasions duringthe past 12 months have you vomited as a result of drinking?

never 0once 2twice 5several or more 8

TOTAL SCORE: Questions 1-11:

A-16

Alcohol Questionnaire-Page 4

SECOND PART

12. Have you ever been told that you have Icohol-related kidney disorders,liver trouble, or cirrhosis? Yes (I) No (0)

13. Have you ever had Delirium Tremers, severe shaking, or hallucinations?Yes (5) (No (0)

14. Have you ever awakened the morning after drinking and found you couldnot recall a part of the evening? Yes (1) No (0)

15a. Have you ever attended a meeting of Alcoholics Anonymous (AA) becauseof your own drinking? Yes _() No (0)

15b. If No, has anyone ever seriously recommended that you attend suchmeetings? Yes (1) No (0)

16. Have you ever seen a clergyman, social worker, doctor, etc., for he.pwith a problem related to your drinking? Yes (1) No (0)

17. Have you ever been in a hospital because of your drinking?Yes (1) No (0)

18. Have you ever been arrested for "drunk and disorderly" or "publicintoxication?" Yes (1) No (0) If Yes, how many times? (x2)

19. Have you ever been arrested for "drunk driving," "driving whileintoxicated," or "driving while under the influence of alcoholic beverages?"Yes (1) No (0) If Yes, how many times? (x2)

TOTAL SCORE: QUESTIONS 12-19:

A-17

Alcohol Questionnaire - Page 5

THIRD PART

20. Do you have any problems with your balance?

Yes _ _ _ _ _No _ _ _ _ _

If yes, describe:

A-18

SCORE

Group A 18+ on Part 11+ on Part 2

or

25+ on Part 10 on Part 2

Group B 18 - 24 Part 10 Part 2

Group C Below 18 on Part 10+ on Part 2

ASSIGNMENTS

1/3 from Group A1/3 from Group B1/3 from Group C

Group A can be assigned to Group.s B & C

Gruup B can be assigned to Groups A & C

Group C can only be assigned to Group C

A-19

Years of Experience in Marine Law Enforcementand OUI Arrest Experience of Officers

Team One Law Enforcement Agency/ Yrs. of Arrests*

Officer # Dept. of Natural Resources Experience

. Ohio 9 35

2 Maryland 14 100

3 Ohio 9 0

Team Two

Officer #

1 Ohio 15 6

2 Ohio 4 20

3 Maryland 15 12

* Arrest - refers to arrests occurring during three boatingseasons, approximately late May to early Sept.1986, 1985, and 1984.

A-20

Item 7

"U- Tru v, t

Mx r

I~~ v.- c~.- 1-1 , S K: t ci;Mix D.11 for S3"-.

> Serve -K' to S3&-2

0)840-58 15br.,a, .3&24 wo.iht ard BT1 result s

0815 Mix DK1 for S3&S4

0820 Serve DK1 to $3&S4

0845 Obtain SI&S2 BT3 results

0845 Mix DK3 for S&S2

0950 Serve DK3 to S3&S2

0900 Obtain S3&S4 BT2 results

0900 Mix DK2 for 53&S4

0905 Serve DK2 to S3&S4

0900-0915 Obtain S5&S6 weight and BT results

0915 Mix DKI for S5&S

0920 Serve DK4 to S5&S

0930 Obtain S1&S2 BT4 results

0930 Mix DK4 for S1&S2

0935 Serve DK4 to S1&S2 for transport to boat..

u945 Obtain $.3&S4 BT3 results

0j45 Mix DK3 for S3&S4

1)950 Serve DK3 tn S3&S4

100 Obtain ;5&S6 BT2 results

A-21

I Mi x FbK f r

1U1.U Mix ;;K4 t r ZT.i&;-4

103 0 M i x ;. 4 f~t :-,- ',4

1()35 - rv& r1K4 t,,-) 'W,",,V14 fr transport to onat.

1045 Ubtain $5&$6 RT3 results

1045 Mix DK3 for S5&S6

1050 Serve DK3 to S5&S6

1100 Obtain S7&S8 BT2 results

i100 Mix DK2 for S7&$8

1105 Serve DK2 to S7&S8

1100-1115 Obtain S9&S10 weight and BT1 results

1115 Mix DK1 for S9&S10

1120 Serve DKI to S9&S10

1130 Obtain S5&S6 BT4 results

1130 Mix DK4 for S5&S6

1135 Serve DK4 to S5&S6 for transport to boat...

1145 Obtain S7&S8 BT3 results

1145 Mix DK3 for S7&S8

1150 Serve DK3 to S7&S8

1200 Obtain S9&SIO RT2 results

1200 Mix DK2 for S9&S10

1205 Serve I)K2 t.o S9&SIO

1200-1215 Ohtain :11&S12 weight and BTI resultsA-22

1:: 1r Mix >.'. f r mr ',-1.i I-i"

1 '35 - ,?4 t, . ,/';.i t,-r tn nsncr. ti 1 h,-,at.

1245 bt ., I , 1w PTJ r 1

1245 M ix i)K: fmr :-V , 1I

1250 Servt [1K3 t. Sg&S 9 0

1300 ()btain S11&S2 BT2 result, s

1300 Mix DK2 for Sl1&S12

1305 Serve DK2 to S11&S12

1330 Obtain S9&S10 BT4 results

1330 Mix DK4 for S9&S10

1335 Serve DK4 to S9&S1O for transport to boat...

1345 Obtain S11&S12 BT3 results

1345 Mix DK3 for SII&S12

1350 Serve DK3 to S1I&S12

1430 Obtain SI1&S12 BT4 results

1430 Mix DK4 for SI1&S12

1435 Serve DK4 to S11&s12 for transport to boat...

Note: Obtain BT5 Results for each pair hourly starting at noon tohave feedback on dosing accuracy.

A-23

Item 8

N M I N A 1, N~ IN [,V F[.-

'4 I ;H T ; ', , i P, , ). i - A

Mi i Mr. M li]i IS : 'lt ! P r Dri nk T t.a I

I " ) - H i ) ! 9 .i ;:5. ,1 '3 1 - 140 1! 7 1 )141 - 150 29 116

5 5 1 16 0)q : ' 3 1 t,- 4

16 1 17.3) 33 132171- 180 24 :T6 :35 14018 1- I90 2h 1 1)4 37 148191-200 :2-7 1)8 39 156201-210 28 112 41 164211-220 .30 121) 43 172221-230 31 124 45 180231-240 32 128 47 188241-250 34 136 48 192

A-24

Item 9

I ll [Ii , i. -, '

7 A N c .Th w

; 11 6 9 4,1l',R [ 2{4 1t66

-I 186 161;: 1 C ;_! l8 lb

Tuesday 5/'12"87

] A 174 1 J2-2 A 175 1323-2 B 186 944-2 A 199 1485-2 C 172 166-2 B 199 967-2 A 197 1488-2 B 157 769-2 C Withdrawn for Emergency Leave10-2 C 252 1611-2 No Show

12-2 C 149 16

Wednesday 5/13/871-3 A 183 1402-3 C 173 163-3 B 205 1064-3 A 182 1405-3 C 173 166-3 C 221 167-3 A 174 1328-3 B 193 1039-3 A 198 15010-3 B 162 8611-3 B 160 8512-3 C 215 16

Thursday 5/14/871-4 C 208 162-4 B 178 1003-4 A 140 1134-4 B 139 55

(Note: Subiect 4-4 arrived late and received only drinks 3&4)5- 4 A 192 1566-4 A 207 1607- 4 A 211 1698-4 C 158 169-4 B 195 108

A-25

r 1 1i 4 B 177

- ! B 1.1 7 H 101:46

11 1

- 156A.- 131 1 **

2 - 170 125

11-5 No Show ---

12-5 A 137 116

Monday 5/18/871-6 B 208 1122-6 B 197 1153-6 B 188 1044-6 B 174 965-6 A 233 1886-6 A 174 1477-6 A 202 1648-6 A 173 1459-6 C 152 16

10-6 No Show11-6 C 193 1612-6 C 195 1613-6 C 162 1614-6 B 202 112

Tuesday 5/19/871-7 A 157 1292-7 A 208 1583-7 B 161 994-7 A 227 1805-7 C 174 166-7 B 152 937-7 A 196 1498-7 No Show ---9-7 No Show

10-7 No Show11-7 C 224 1612-7 B 212 125

Wednesday 5/20/871 - A 146 1232 B 195 1083 -8 C 164 16

A-26

J - A 1 1 1>.'

A I:; 4

1 72

:- / A 181 1484" - 1( , 09 149 A 164 1

1 : , ,, ,; i~ w .. .

79B 12 111

8.9 B 167 16

1<: A 24i; 191S: A 18 1 148

9-9 C 176 164-9 c 29 165-9 C 213 166-9 A 162 1397-9 B 182 1118-9 B 167 999-9 B 189 104

10-9 B 179 10211-9 C 207 1612-9 C 191 16

13-9 C 217 16

A-27

ftem 10

MEDICAL REPORT

Name: _Height:

Address: Weight:

Age:

Phone:

Pulse: Blood Pressure: Heart:

Temp: Respiration: Color:

Are you presently taking any prescription or over-the-counter medication ordrugs? Yes No

If yes, what?

Examination Date:

Subject is/is not qualified to participate in controlled drinking study.

Yes No

Examininer:

A-28

Item 11

BRIEFING

Introduction

Iam

As you know, we are requesting volunteers to participate in a research projectto evaluate standard methods for testing intoxication. Today I am going to askyou to fill out a brief questionnaire to help us with our planning. Someone' willcall you in the next day or two to confirm the time and date of yourparticipation. We are planning to start the project on May 11 and run it forabout two weeks so your session will be somewhere during that time.

I will now go through the procedure so that you will understand what will beexpected of you.

At the agreed upon time, you will arrive at the site of the study, where youwill be expected to spend approximately 7 hours. Soon after your arrival, anattending nurse will conduct a brief medical screening. The nurse will checkyour blood pressure, temperature and pulse and ask about any medication youmight be taking. The nurse will remain in attendance throughout the study.You will then be asked to consume a glass of fruit juice with a measured amountof alcohol each 40 minutes until three glasses have been consumed. Before andafter each drink you will be asked to provide a breath sample by blowing intoan automatic alcohol testing device. Following the third drink and breath test,you will be asked to board a boat for a boat trip. During the trip you will beasked tc, consume a 4th drink followed by another breath test. Police officerswill then board the boat and administer standard sobriety tests. At the end ofthis trip, the boat will be docked and boarded by police officers who will giveyou further standard sobriety tests. You will be escorted off the boat. Youwill be given a meal and then asked to provide breath samples for testing atregular intervals to monitor the elimination of alcohol from your body. Whenyour alcohol content level is at acceptable and legal limits for sobriety you willbe transported to your place of residence.

You will be asked to sign an Informed Consent Form, a copy of which has beengiven to you for review.

Are there any questions?

Now you can proceed with completing the questionnaire. Please be assured thatyour responses will be considered private and confidential information and willonly be seen by myself or people from my company. When the project is overit will be destroyed. If you have any questions while you are doing this pleasecome up to see me.

A-29

Item 12

INFORMED CONSENT FORM

Gentlemen:

I have agreed to participate in a research project to evaluate stanaard methodsfor testing intoxication. My consent to participate in this experimental researchproject does not in any way imply or constitute employment with the FederalGovernment or Dunlap and Associates, Inc. (D&A).

I understand the following conditions:

At an agreed upon time, I will arrive at the site of the study, where I shallexpect to spend approximately 8 hours. Soon after my arrival, an attendingnurse will conduct a brief medical screening and will remain in attendancethroughout the study. I will then be asked to consume a glass of fruit juicewith a measured amount of alcohol each 40 minutes until three glasses have beenconsumed. Before and after each drink I will be asked to provide a breathsample by blowing into an automatic alcohol testing device. Following the thirddrink and breath test, I will be asked to board a boat for a boat trip. Duringthe trip I will be asked to consume a 4th drink followed by another breath test.Police officers will then board the boat and administer standard sobriety tests.At the end of this trip, the boat will be docked and boarded by police officerswho will give me further standard sobriety tests. I will be escorted off theboat. I will be given a meal and then asked to provide breath samples fortesting at regular intervals to monitor the elimination of alcohol from my body.When my alcohol content level is at acceptable and legal limits for sobriety I willbe transported to my place of residence.

There has been no coercion, element of fraud or deceit, undue moral suasion orother adverse pressure brought to bear in my volunteering for this study. Ihave done so of my own free will and I can request that my participation beterminated at any time if in my opinion I have reached-the physical or mentalstate where continuation becomes undesirable.

I air over the age of 21 years and I have no known physical or medicalinfirmities nor am I presently using any medication or drugs which make myparticipation in this program dangerous to my well-being or to that of others.I also understand that any information I provide to the study personnel will bekept in total confidence and will only be used anonymously in statisticalcombination with information from other participants and that the AlcoholQuestionnaire that I completed will be destroyed at the end of the study.

My relationship to the Federal Government and D&A is that of a volunteerparticipant in a research project and nothing contained herein shall beconstrued as creating any other relationship. I hereby release and holdharmless DaA, their respective agents, servants and employees from any and allclaims arising from my participation in or the conduct of the management of theprogram herein contemplated.

Witness Signature

Date

Address

Ieiepnone No.

A-30

References

Burns, MI. and Moskowitz, H. 'Psychophysical Tests for DWVI Arrest". (Contract No.DOT-1-HS-5-O! 2421). Washington, D.C.: Department of Transportation, NHfTSA, June 1977.

Yharp, V., Burns, M., a nd Moskowvitz, H . 'Development o f Field Test 0 fPsvc hophysical Tests for DXVI Arrest", (Contract No. DOT-HS-8-01970). Washington,D.C.: Department o~f Transportation, NHTSA, Mvarch 1981.

R-1/R-2